Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Cloud-Based Breast Cancer Prediction Empowered with Soft Computing Approaches

Cloud-Based Breast Cancer Prediction Empowered with Soft Computing Approaches Hindawi Journal of Healthcare Engineering Volume 2020, Article ID 8017496, 16 pages https://doi.org/10.1155/2020/8017496 Research Article Cloud-Based Breast Cancer Prediction Empowered with Soft Computing Approaches 1,2 1,3 1 4 Farrukh Khan, Muhammad Adnan Khan , Sagheer Abbas , Atifa Athar, 1,5 1,5 6 Shahan Yamin Siddiqui, Abdul Hannan Khan, Muhammad Anwaar Saeed, 1,2 and Muhammad Hussain Department of Computer Science, National College of Business Administration and Economics, Lahore, Pakistan Department of Computer Science, Lahore Institute of Science and Technology, Lahore, Pakistan Department of Computer Science, Lahore Garrison University, Lahore, Pakistan Department of Computer Science, CUI, Lahore Campus, Pakistan Department of Computer Science, Minhaj University, Lahore, Pakistan Department of Computer Science, Virtual University, Islamabad, Pakistan Correspondence should be addressed to Muhammad Adnan Khan; madnankhan@ncbae.edu.pk Received 18 January 2020; Accepted 30 April 2020; Published 19 May 2020 Academic Editor: Norio Iriguchi Copyright © 2020 Farrukh Khan et al. )is is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. )e developing countries are still starving for the betterment of health sector. )e disease commonly found among the women is breast cancer, and past researches have proven results that if the cancer is detected at a very early stage, the chances to overcome the disease are higher than the disease treated or detected at a later stage. )is article proposed cloud-based intelligent BCP-T1F- SVM with 2 variations/models like BCP-T1F and BCP-SVM. )e proposed BCP-T1F-SVM system has employed two main soft computing algorithms. )e proposed BCP-T1F-SVM expert system specifically defines the stage and the type of cancer a person is suffering from. Expert system will elaborate the grievous stages of the cancer, to which extent a patient has suffered. )e proposed BCP-SVM gives the higher precision of the proposed breast cancer detection model. In the limelight of breast cancer, the proposed BCP-T1F-SVM expert system gives out the higher precision rate. )e proposed BCP-T1F expert system is being employed in the diagnosis of breast cancer at an initial stage. Taking different stages of cancer into account, breast cancer is being dealt by BCP-T1F expert system. )e calculations and the evaluation done in this research have revealed that BCP-SVM is better than BCP-T1F. )e BCP-T1F concludes out the 96.56 percentage accuracy, whereas the BCP-SVM gives accuracy of 97.06 percentage. )e above unleashed research is wrapped up with the conclusion that BCP-SVM is better than the BCP-T1F. )e opinions have been recommended by the medical expertise of Sheikh Zayed Hospital Lahore, Pakistan, and Cavan General Hospital, Lisdaran, Cavan, Ireland. formed in the breast cells; when these cells begin to grow 1.Introduction irregularly in a human body, it results in flaking and redness Apparently, the diagnosis and the scrutiny of the breast of the breast. )e cancer is still considered as the undiag- cancer disease have always been a decisive and critical one in nosed and untreated disease in various parts of the world. the regard of medical department. )e cancerous lumps )e questions arise, why the cancer still has the strong roots form in a particular area of the body when the human cells in patients? Why still breast cancer remains undiagnosed begin to produce rapidly beyond the expected limit. )e amongst the women? Why the ratio of mortality among cancerous lumps which are also termed as tumors are women due to breast cancer is still constant? )e breast comprised of two kinds: one is the benign and the other is cancer should be diagnosed at an early stage so that the malignant. Breast cancer is considered as a lump that is condition does not persist and many lives could be saved. If 2 Journal of Healthcare Engineering it gets diagnosed initially, then the chances to overcome the 2. Literature Review disease are certainly higher. Women are being impinged by To avoid the faulty reasoning processes, errors, continuous this disease commonly. So if cancer remains undiagnosed, failures, lack of knowledge, and failures in rules of logics in then it may lead to death [1, 2]. )e risk factors evolve detection of tumors, researchers have started to advance new because of which breast cancer is induced may be the genetic methodologies, models, and tools. Bearing a clear aim, reasons, alcohol intake, dense tissues in breast, radiation various systems and proposed models have been put forward exposure, age, and so on. Since 1989, the survival rate in successful identification of breast cancer. Among women amongst masses has been immensely improved due to in America at the rate of one in three cancers approximately, modern technologies introduced in screening and treat- breast cancer is the most frequently diagnosed cancer among ment. )e recently conducted researches have shown that, in the women [6, 12, 13]. Surgical biopsies confirm malignancy 2017, 252,710 women were diagnosed with this disease; with high level of sensitivity but are considered costly and approximately 40,610 according to the statistics were more can affect patient’s psychology as well. )is research dem- likely to die from breast cancer. )e radical steps in reducing onstrates novel approach by using morphological operators the risk factors of this disease are the awareness of the in- and clustering algorithm fuzzy c-m to identify malignant duction causes of the breast cancer. )e early symptoms of lump in mammography automatically [6, 14–17]. In the cancer and screening can lessen the factors of cancer in- article, initial identification of tumor, fuzzy system’s various fringing masses [3]. applications, and algorithms have been proposed [18, 19]. In )e physical diagnosis of this mortal disease involves some previous studies on FDTs, proposed approaches focus the breast exam, imaging tests, biopsy, blood tests, and so on modification of decision tree pruning algorithm and on. )e initial blood marker tests which are CA 15.3, require fuzzy parameters to be set by domain experts. We TRU-QUANT, CA 125, CEA, and so on are done before opted to fuzzify already generated decision tree nodes to the treatment of this cancer. )e blood marker tests work relax the sharp decision boundaries. A similar kind of ap- as an initial indicator for this disease; in determining the proach is employed in [20, 21]. breast cancer, currently there are four main methods Fuzzy logic has been rarely used in cancer prognosis. which are employed in differentiating between the ma- Being noncrisp, it can act as a natural ally of a physician in lignant lump and benign lump: biopsy, fine needle as- prognostic decision-making process [22]. In the recent re- piration, mammography, and MRI [4, 5]. )e proposed searches, we have surveyed various types of research sce- methodology has employed the two algorithms to detect narios [21–23]; in the prognosis of cancer, the applications of an ailment efficiently. )e Mamdani inference system is various cases of machine learning techniques are contrib- used among the recent researches to detect a particular uting towards the advanced researches. Some of the basic kind of disease; this proposed research has compiled the trends which are encountered for the motivation of ex- previously done evaluations in comparison to this pro- periments include the following: the fuzzy logic has been posed methodology [6]. Mammography determines the used in the diagnosis of cancer rarely. Aiming for clear incomplete diagnosis of the infection; if the infection interpretation of a particular type of disease physicians, results conclude out to be negative, the infection turns out using “Black Box” models, approximately 70% of all re- as benign. Mammography determines whether there is searches reported making use of neural networks. )e any lump, no lump, or cancerous lump. Mammography majority of the manuscripts used machine learning tech- concludes the severity level and the type of cancer is niques independently without considering potential in the measured with the help of biopsy gold test. )e biopsy discussed manuscripts to cope up with each other in a hybrid gold test (type) will also determine the benign, ductal model. Lack of attention is paid to data size. Victor gives one carcinoma, invasive lobular carcinoma, inflammatory solution to computerized tool used for diagnosis of breast breast disease, and lobular carcinoma [7, 8]. cancer. )e fact is that fuzzy logic can substantially assist in )e recent researches show that, in 2017, 252,710 women diagnosis of breast cancer is being put forward in this paper were diagnosed with this disease; approximately 40,610 [24–26]. which were more likely to die from breast cancer [1, 9, 10]. Diagnosis of breast cancer is through fuzzy clustering )e physical diagnosis involves the breast exam, imaging with partial supervision [27]. ARTMAP approach gives tests, biopsy, blood tests, and so on. In regard of determining 97.2% accuracy, which represents the one-way approach in the breast cancer, currently there are four main methods the neural networks for the diagnosis of breast cancer [28]. which are used in differentiating between the malignant Classification exactness of over 95% was professed to be lump and benign lump: biopsy, fine needle aspiration, accomplished by utilizing little MIAS database; the proposed mammography, and MRI [2, 3, 11]. framework is being acknowledged for the findings of bosom In this research, fuzzy logic and SVM are used to find out malignancy dependent on outrageous learning machine the breast cancer by using different statistical measures such [29–32]. as accuracy, miss rate, specificity, sensitivity, false-positive Resisting a technique for classification of mammogram value, false-negative value, likelihood ratio positive, likeli- that is comprised of 4 phases, preprocessing stage utilized hood ratio negative, positive prediction value, and negative middle filter to upgrade nature of picture and to expel prediction values. With the help of these matrices, breast clamor from the picture. To check the variation from the cancer can be found more accurately as compared with the norm of the mammograms, ANN classifier was utilized to previous literature. Journal of Healthcare Engineering 3 group the picture into fitting class. Affectability, specificity, defined easily by the help of a medical expert. Type-1 fuzzy and precision asserted in the work were 72.72%, 93.6%, and logic rules are applied on inputs of fuzzy sets and then converted it into a fuzzy output. In this research, input 88.66% [18, 33]. A framework for the findings of bosom malignancy variables are used to propose a system to diagnose the dependent through feedforward networks was proposed. particular disease which is cardiac by using a fuzzy logic Prepreparing was done in two-phase foundation and second model. For the detection of cardiac disease, Support Vector was evacuating pectoral muscle. Hough change strategy was Machine (SVM) is a model that provides computational utilized for ROI. An aggregate of 32 dark dimensions and results which depend upon the structure and biological surface highlights is separated from mammograms. Preci- functions of neural networks. In the prediction layer, sion asserted by utilizing smaller than normal MIAS data- Support Vector Machine is used to find out the breast base was 94.06% [34]. cancer, and in the performance layer, it is used to evaluate A few systems have been sent to anticipate and perceive the results produced by the prediction layer performance. significant example for breast malignancy analysis. Data )e whole system process is shown in Figure 1, in which the data acquisition layer comprises the parameters of input. In mining further categorizes the different methods of the decision tree, ANN, RIPPERS classifier, and Support Vector this model, they will go for the neural system, where a Machine (SVM) to make a quick explanation and survey of trained algorithm is used to estimate breast cancer. At the the dataset regarding the cardiovascular disease. )e ex- industrial level, SVM is utilized and it gives accurate results. planation used the consideration and comparison of the SVM includes several neurons that are specifically orga- performance of the techniques which encompasses accuracy, nized. Neurons and influences among them are essential sensitivity, specificity, error rate, true positive rate, and false parts of an SVM. Neurons have handling features that co- positive rate [35, 36]. operate to overcome an issue. )is layer is used to examine Computational intelligence approaches like fuzzy system the breast cancer on the basis of thirty input parameters, [37–39], neural network [40], and swarm intelligence [41] which is termed as the scientific study of the models that are and evolutionary computing [42] like genetic algorithm statistical in nature and constitute algorithms which com- puter systems employ to perform a certain type of task with [43, 44], DE, Island GA [45], Island DE [46, 47], classifier [48], and SVM [49] are strong candidate solutions in the greater precision and as certainty. In the performance field of smart city [50], wireless communication, and so on. evaluation layer, precision and miss rate are determined. In the decisive area, the conclusion is made whether the breast 3.Proposed System Model Methodology cancer is identified or not. )e following methodology has been elaborated in Figure 1. First layer is the data acquisition layer which follows up with 3.1. Fuzzy System Methodology. Our proposed model breast the data collection of breast cancer. )e raw data attained cancer prediction (BCP), multilayered Mamdani fuzzy type- through the collection of breast cancer is then fed into the 1 inference system- (MFIS-) based expert system (BCP-T1F) preprocessing layer. )e preprocessing layer is a criterion to is explained in this section. )e BCP-T1F expert system consists of four layers as shown in Figure 2. In layer 1 named handle the missing values amongst the raw data; further- more, moving average and normalizations are being done in symptoms, the initial symptoms of breast cancer will be checked which are swelling, breast pain, redness, nipple the preprocessing layer. )e omissions and errors are being lessened through the standard portable. )en after the retraction, Family Inheritance Breast Cancer (FBIC), and completion of the previous layers, the preprocessing layer skin irritation. )is will find whether it is lump or cancer. then jumps on to the application layer. )e application layer If the system finds the symptoms in the patient, then is comprised of the prediction layer and the performance layer 2 diagnoses the breast cancer (no/yes) using two input evaluation layer. )e prediction layer specifically focuses on variables that are ultrasound and mammography. If the layer the two algorithms which are employed to determine the 2 diagnoses breast cancer, then layer 3 will be activated. Layer 3 predicts the type and severity of BCP based on two indispensable types of breast cancer through type-1 fuzzy logic and SVM just points out that something is fishy or not; input variables (biopsy gold severity) and (biopsy gold type). )en layer 4 will check the stage of cancer by three input that is, a person is suffering from the disease or not. )e two algorithms which aided the application layer are shown in variables that are MRI, CT, and PET which are shown in Figure 2. Figure 1. Type-1 fuzzy logic is an enabled system used to get accurate results from big data. )e performance evaluation Mathematically, the layers of the proposed BCP-T1F can layer calculates the accuracy and miss rate. Type-1 fuzzy be written as follows. logic constitutes of logical rules and these rules can be )is layer 1 can be written mathematically as μ � MFIS􏽨μ , μ , μ , μ , μ , μ 􏽩. (1) DBI,Layer1 swelling Breast pain Redness Skin irritation FBIC nipple−retraction 4 Journal of Healthcare Engineering Application layer Data Pre- Performance Prediction acquisition processing evaluation layer layer layer layer Handling missing Type-1 Malignancy Is breast value Refer to fuzzy cancer emergency logic Accuracy identified? Breast cancer Moving data average Raw collection data Miss rate Support Benign vector machine Normali- zations Discard Figure 1: Proposed intelligent breast cancer prediction model for BCP-T1F-SVM expert system. Biopsy gold type MRI Ultrasound Skin redness CT Biopsy gold type Mammography PET Lump/cancerous lump Mild-grad/High Breast pain severity/low severity Infected BCP Condition Condition Layer 3 Layer 4 Condition Redness Layer 1 Layer 2 No cancerous Normal No lump FBIC lump Discard Discard Discard Nipple retraction Figure 2: Proposed BCP-T1F expert system methodology. )e layer 2 can be written as )en layer 4 can be written as μ � MFIS μ , μ . (2) μ � MFIS􏼂μ , μ , μ 􏼃. (4) 􏽨 􏽩 DBC−ST, Layer 4 MRI CT PET DBI−RADS,Layer 2 Ultrasond Mammography )en layer 3 can be written as 3.1.1. Membership Functions. )e membership function of μ � MFIS􏽨μ , μ 􏽩. DBC−ST, Layer 3 Biopsy Gold−Type Biopsy Gold−Severity proposed BCP-T1F expert system yields the curve values (3) ranging between 0 and 1 and also dispenses a mathematical Journal of Healthcare Engineering 5 Table 1: Input and output variables membership functions used in the proposed BCP-T1F expert system. Sr. I/P parameters Mathematics of membership function number μ (α) � {max(min(1, ((0.5 − α)/0.1)), 0)}, Swelling−affected 1 Swelling (μ (α)) Swelling μ (α) � max(min(1, ((α − 0.4)/0.1) ), 0) { } Swelling−not affected μ (β) � 􏼈max(min(1, ((1.5 − β)/0.1)), 0)􏼉, Skin−irritation,affected 2 Skin irritation (μ (β)) Skin−irritation μ (β) � 􏼈max(min(1, ((β − 1.4)/0.1)), 0)􏼉 Skin−irritation,not affected μ (c) � 􏼈max(min(1, ((2.5 − c)/0.1)), 0)􏼉, Breast pain−affected 3 Breast pain (μ (c)) Breast−pain μ (c) � 􏼈max(min(1, ((c − 2.4)/0.1)), 0)􏼉 Breast pain−not affected μ (ρ) � max(min(1, ((3.5 − ρ)/0.1)), 0) , 􏼈 􏼉 Redness−affected 4 Redness (μ (ρ)) Redness μ (ρ) � 􏼈max(min(1, ((ρ − 3.4)/0.1)), 0)􏼉 Redness−not affected Family inheritance, 5 μ (ψ) � 􏼈max(min(1, ((4.5 − ψ)/0.1)), 0)􏼉, μ (ψ) � 􏼈max(min(1, ((ψ − 4.4)/0.1)), 0)􏼉 E,N E,P Breast cancer (μ (ψ)) FIBC Nipple retraction μ (φ) � 􏼈max(min(1, ((5.5 − φ)/0.1)), 0)􏼉, Nipple retraction−affected (μ (φ)) μ (φ) � max(min(1, ((φ − 5.4)/0.1)), 0) 􏼈 􏼉 Nipple−retraction Nipple retraction−not affected 1, if di € 􏼂 0 0.4 􏼃, ⎧ ⎪ ⎫ ⎪ ⎨ ⎬ μ (di) � ((0.5 − di)/0.1), if di€ 􏼂 0.4 0.5 􏼃, , DI−normal ⎪ ⎪ ⎩ ⎭ 0, otherwise 7 Diagnosis infection (μ (di)) DI ⎧ ⎪ ((di − 0.4)/0.1), if di€ 􏼂 0.4 0.5 􏼃 ⎫ ⎪ ⎨ ⎬ μ (di) � 1, if di€ 0.5 1 􏼂 􏼃 DI−infected ⎪ ⎪ ⎩ ⎭ 0, otherwise Table 2: Input and output variables membership functions used in the proposed BCP-T1F expert system. Sr. I/P parameters Mathematics of membership function number μ (λ) � {max(min(1, ((0.5 − λ)/0.1)), 0)}, Swelling−affected 1 Ultrasound (μ (λ)) Ultrasound μ (λ) � max(min(1, ((λ − 0.4)/0.1)), 0) { } Swelling−not affected μ (τ ) � 􏼈max(min(1, ((0.9 − τ )/0.1)), 0)􏼉, Mammography−incomplete 1 1 μ (τ ) � 􏼈max(min(((τ − 0.8)/0.1), 1, ((1.9 − τ )/0.1)))􏼉, Mammography−negative 2 2 2 μ (τ ) � 􏼈max(min(((τ − 1.8)/0.1), 1, ((2.9 − τ )/0.1)))􏼉, Mammography−benign 3 3 3 2 Mammography (μ (τ)) μ (τ ) � 􏼈max(min(((τ − 2.8)/0.1), 1, ((3.9 − τ )/0.1)))􏼉, Mammography Mammography−probably benign 4 4 4 μ (τ ) � max(min(((τ − 3.8)/0.1), 1, ((4.9 − τ )/0.1) )) , 􏼈 􏼉 Mammography−suspicious 5 5 5 μ (τ ) � 􏼈max(min(((τ − 4.8)/0.1), 1, ((5.9 − τ )/0.1)))􏼉, Mammography−suggested malignancy 6 6 6 μ (τ ) � max(min(1, ((τ − 5.8)/0.1), 0)) 􏼈 􏼉 Mammography−proven malignnancy 7 7 1, if rads€ 0 0.3 ⎧ ⎪ 􏼂 􏼃 ⎫ ⎪ ⎨ ⎬ μ (rads) � ((0.4 − rads)/0.1) if rads€ 􏼂 0.3 0.4 􏼃 , DBI−Rads−No lump ⎪ ⎪ ⎩ ⎭ 0, otherwise ((rads − 0.3)/0.1), if rads€ 􏼂 0.3 0.4 􏼃 ⎧ ⎪ ⎫ ⎪ ⎪ ⎪ Detection of breast imaging reporting ⎪ ⎪ ⎨ ⎬ 1, if rads€ 􏼂 0.4 0.6 􏼃 3 and database system score μ (rads) � , DBI−Rads−Lump ⎪ ⎪ ((0.7 − rads)/0.1) if rads 0.6 0.7 ⎪ 􏼂 􏼃 ⎪ ⎪ ⎪ (μ (rads)) ⎩ ⎭ DBI−rads 0, otherwise 1, if rads€ 􏼂 0.7 1 􏼃 ⎧ ⎪ ⎫ ⎪ ⎨ ⎬ μ (rads) � ((rads − 0.6)/0.1) if rads€ 􏼂 0.6 0.7 􏼃 DBI−Rads−Cancerous lump ⎪ ⎪ ⎩ ⎭ 0, otherwise form of the fuzzy logic that accords statistical values of both 3.1.2. Rules Table. )e proposed system BCP-T1F rules table the input and output variables. )e mathematical repre- usually relies on the expert system which constitutes of sixty- sentation of proposed BCP-T1F expert system yields four inputs and output rules for layer 1, fifteen output and member functions of layers 1–4 shown in Tables 1–4. )ese input rules for layer 2, eight input and output rules for layer membership functions are gathered after the consultation 3, and thirty I/O rules for layer 4. )is rule (Table 5) has been with the medical experts from Cavan General Hospital obtained with the assistance of the medical experts from Lisdaran, Cavan, Ireland. Cavan General Hospital Lisdaran, Cavan, Ireland. 6 Journal of Healthcare Engineering Table 3: Input and output variables membership functions used in the proposed BCP-T1F expert system. Sr. number I/P parameters Mathematics of membership function μ (δ ) � 􏼈max(min(1, ((5 − δ )/1)), 0)􏼉, Biopsy gold−severity−benign 1 1 Biopsy gold standard for severity 1 μ (δ ) � max(min(((δ − 4)/1), 1, ((10 − δ )/1)), 0) , 􏼈 􏼉 Biopsy gold−severity−cancer−severity low 2 2 2 (μ (δ)) Biopsy gold−severity μ (δ ) � 􏼈max(min(1, ((δ − 9)/1)), 0)􏼉 Biopsy gold−severity−cancer−severity high 3 3 μ (π ) � 􏼈max(min(1, ((5 − π )/1)), 0)􏼉, Biopsy gold−type−fine needle aspiration 1 1 Biopsy gold standard for type 2 μ (π ) � 􏼈max(min(((π − 4)/1), 1, ((10 − π )/1)), 0)􏼉, Biopsy gold−type−core needle biopsy 2 2 2 (μ (π)) Biopsy gold−type μ (π ) � 􏼈max(min(1, ((π − 9)/1)), 0)􏼉 Biopsy gold−type−surgical biopsy 3 3 1, if ∈ 􏼂 0 0.3 􏼃 ⎧ ⎪ ⎫ ⎪ ⎨ ⎬ μ (ρ ) � ((0.4 − ρ )/0.1), if ∈ 􏼂 0.3 0.4 􏼃 , DBCS−mild grade 1 1 ⎪ ⎪ ⎩ ⎭ 0, otherwise ((ρ − 0.3)/0.1), if ∈ 0.3 0.4 ⎧ ⎪ 􏼂 􏼃 ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ Diagnosis of breast cancer severity 1, if ∈ 􏼂 0.4 0.6 􏼃 3 μ (ρ ) � , DBCS−moderate severity 2 ⎪ ⎪ (μ (ρ)) ⎪ ((0.7 − ρ )/0.1), if ∈ 􏼂 0.6 0.7 􏼃 ⎪ DBCS 2 ⎪ ⎪ ⎩ ⎭ 0, otherwise ⎧ ⎪ 1, if rads ∈ 􏼂 0.7 1 􏼃 ⎫ ⎪ ⎨ ⎬ μ (ρ ) � ((ρ − 0.6)/0.1), if rads ∈ 0.6 0.7 􏼂 􏼃 DBCS−high severity 3 ⎪ 3 ⎪ ⎩ ⎭ 0, otherwise ⎧ ⎪ 1, if ∈ 􏼂 0 0.2 􏼃 ⎫ ⎪ ⎨ ⎬ μ (b ) � ((0.3 − b , )/0.1) if ∈ 􏼂 0.2 0.3 􏼃 DBCT−severity 1 1 ⎪ ⎪ ⎩ ⎭ 0, otherwise ((b − 0.2)/0.1), if ∈ 􏼂 0.2 0.3 􏼃 ⎧ ⎪ ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ 1, if ∈ 􏼂 0.3 0.4 􏼃 μ (b ) � , DBCT−ductal carcinoma 2 ⎪ ⎪ ⎪ ((0.5 − b )/0.1) if ∈ 􏼂 0.4 0.5 􏼃 ⎪ ⎪ ⎪ ⎩ ⎭ 0, otherwise 1, if ∈ 􏼂 0.5 0.6 􏼃 ⎧ ⎪ ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ Diagnosis of breast cancer ((b − 0.4)/0.1), if ∈ 􏼂 0.4 0.5 􏼃 4 μ (b ) � , DBCT−invasive lobular carcinoma 3 ⎪ ⎪ type (μ (b)) ⎪ ((0.7 − b )/0.1), if ∈ 􏼂 0.6 0.7 􏼃 ⎪ DBCT ⎪ 3 ⎪ ⎩ ⎭ 0, otherwise 1, if ∈ 0.6 0.7 ⎧ ⎪ 􏼂 􏼃 ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ ((b − 0.6)/0.1), if ∈ 􏼂 0.7 0.8 􏼃 μ (b ) � , DBCT−inflammatory breast disease 4 ⎪ ⎪ ⎪ ((0.9 − b )/0.1), if ∈ 􏼂 0.8 0.9 􏼃 ⎪ ⎪ ⎪ ⎩ ⎭ 0, otherwise ⎧ ⎪ 1, if ∈ 􏼂 0.9 1 􏼃 ⎫ ⎪ ⎨ ⎬ μ (b ) � ((b − 0.8)/0.1), if ∈ 0.8 0.9 􏼂 􏼃 DBCT−lobular carcinoma 5 ⎪ 5 ⎪ ⎩ ⎭ 0, otherwise 3.1.3. Rule Based. Rules are essential for input and output 3.1.6. Lookup Diagram. MATLAB R2019a tool is used for variables. Achievement of an adroit system is built based on demonstrating, imitation, algorithm expansion, prototyping, rules. Some of the rules are shown in Table 4. and various other fields. )is tool is well organized for software designing, data examination, conception, and calculations. For the simulation of results, three inputs and one output of BCP 3.1.4. Inference Engine. Inference engine is the most em- are used on layer 4 which are shown in Figure 4. phasized constituent of any decision-based expert system. In Figure 4 shows that (μ (o)) is considered as CT nodes mode this manuscript, BCP-T1F expert system has been employed node 2, (μ (])) is taken into account as low size, MRI−tumor size in layer 1, layer 2, layer 3, and layer 4. (μ (θ)) is spread in whole body, and (μ (χ)) turns PET DBC−stage out to be concluded as Stage 3. Similarly, Figure 4 also demonstrates the rule-based 3.1.5. Defuzzification. Defuzzification is the process of making knowledge; few of them are shown as follows: a measureable result in crusty logic, given fuzzy sets, and (μ (o)) is considered as node 3, CT nodes mode corresponding membership degrees. It is the process that plots a (μ (])) is taken into account as very high size, MRI−tumor size fuzzy set to a crisp set. It is characteristically needed in fuzzy (μ (θ)) is spread in whole body, and (μ (χ)) PET−benign DBC−stage control systems. In Figures 3(a)–3(d), the graphical illustrations turns out to be concluded as Stage 4. of defuzzifier of BCP-T1F expert system are obtainable. Journal of Healthcare Engineering 7 Table 4: Input and output variables membership functions used in the proposed BCP-T1F expert system. Sr. I/P parameters Mathematics of membership function number ⎧ ⎪ 1, if o ∈ 􏼂 0 7.9 􏼃 ⎫ ⎪ ⎨ ⎬ (μ (o)) � ((10 − o)/(10 − 9.9)), if o ∈ 9.9 10 , 􏼂 􏼃 CT nodes − mode 1 ⎪ ⎪ ⎩ ⎭ 0, otherwise ((o − 9.9) /(10 − 9.9)), if o ∈ 􏼂 9.9 10 􏼃 ⎧ ⎪ ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ Computed tomography (CT) 1, if o ∈ 􏼂 10 19.9 􏼃 1 (μ (o )) � , CT nodes − mode 1 2 ⎪ ⎪ (μ (o)) ⎪ ((20 − o) /(20 − 9.9)), if o ∈ 􏼂 19.9 20 􏼃 ⎪ CT nodes mode ⎪ ⎪ ⎩ ⎭ 0, otherwise ⎧ ⎪ ((o − 19.9)/(20 − 19.9)), if o ∈ 􏼂 19.9 20 􏼃 ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ 1, if o ∈ 􏼂 20 30 􏼃 (μ (o )) � CT nodes − mode 1 3 ⎪ ⎪ ⎪ ((20 − o) /(20 − 9.9)), if o ∈ 􏼂 19.9 20 􏼃 ⎪ ⎪ ⎪ ⎩ ⎭ 0, otherwise 1, if ] ∈ 0.02 0.1 ⎧ ⎪ 􏼂 􏼃 ⎫ ⎪ ⎨ ⎬ μ (]) � ((0.2 − ])/0.1) if ] ∈ 􏼂 0.1 0.2 􏼃 , MRI − tumor size− no ⎪ ⎪ ⎩ ⎭ 0, otherwise ((] − 0.1)/0.1), if ] ∈ 􏼂 0.1 0.2 􏼃 ⎧ ⎪ ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ 1, if ] ∈ 􏼂 0.2 1.9 􏼃 μ (]) � , MRI−tumor size−low ⎪ ⎪ ((2 − ])/0.1), if ] ∈ 1.9 2 ⎪ 􏼂 􏼃 ⎪ ⎪ ⎪ ⎩ ⎭ 0, otherwise 1, if ] ∈ 􏼂 2 4.9 􏼃 ⎧ ⎪ ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ Magnetic resonance imaging ((] − 1.9)/0.1), if ] ∈ 􏼂 1.9 2 􏼃 2 μ (]) � , MRI − tumor size − medium ⎪ ⎪ (μ (])) ⎪ ((5 − ])/0.1), if ] ∈ 􏼂 4.9 5 􏼃 ⎪ MRI−tumor size ⎪ ⎪ ⎩ ⎭ 0, otherwise ⎧ ⎪ 1, if ] ∈ 􏼂 5 7.9 􏼃 ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ ((] − 4.9)/0.1), if ] ∈ 􏼂 4.9 5 􏼃 μ (]) � , MRI − tumor size − high ⎪ ⎪ ⎪ ((8 − ])/0.1), if ] ∈ 􏼂 7.9 8 􏼃 ⎪ ⎪ ⎪ ⎩ ⎭ 0, otherwise ⎧ ⎪ 1, if ] ∈ 􏼂 8 11 􏼃 ⎫ ⎪ ⎨ ⎬ μ (]) � ((] − 7.9)/0.1) if ] ∈ 􏼂 7.9 8 􏼃 DBCT − lobular carcinoma ⎪ ⎪ ⎩ ⎭ 0, otherwise 1, if θ ∈ 0 0.4 ⎧ ⎪ 􏼂 􏼃 ⎫ ⎪ ⎨ ⎬ μ (θ) � ((0.5 − θ)/0.1) if θ ∈ 􏼂 0.4 0.5 􏼃 , PET − benign ⎪ ⎪ ⎩ ⎭ 0, otherwise Positron emission tomography (μ (θ)) PET ((θ − 0.4)/(0.5 − 0.4)), if θ ∈ 􏼂 0.4 0.5 􏼃 ⎧ ⎪ ⎫ ⎪ ⎨ ⎬ μ (θ) � 1, if θ ∈ 􏼂 0.5 1 􏼃 PET _in_whole_body ⎪ ⎪ spread ⎩ ⎭ 0, otherwise 8 Journal of Healthcare Engineering Table 4: Continued. Sr. I/P parameters Mathematics of membership function number ⎧ ⎪ 1, if χ ∈ 􏼂 0 0.2 􏼃 ⎫ ⎪ ⎨ ⎬ (μ (χ)) � , ((0.3 − χ)/(0.3 − 0.2)), if χ ∈ 􏼂 0.2 0.3 􏼃 DBC−stage 1 ⎪ ⎪ ⎩ ⎭ 0, otherwise ((χ − 0.2) ( / 0.3 − 0.2)), if χ ∈ 􏼂 0.2 0.3 􏼃 ⎧ ⎪ ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ 1, if χ ∈ 􏼂 0.3 0.4 􏼃 (μ (χ)) � , DBC−stage 2 ⎪ ⎪ ⎪ ((0.5 − χ) /(0.5 − 0.4)), if χ ∈ 􏼂 0.4 0.5 􏼃 ⎪ ⎪ ⎪ ⎩ ⎭ 0, otherwise ⎧ ⎪ ((χ − 0.4) /(0.5 − 0.4)), if χ ∈ 􏼂 0.4 0.5 􏼃 ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ Diagnosis of breast cancer 1, if χ ∈ 􏼂 0.5 0.6 􏼃 4 (μ (χ)) � , DBC−stage 3 ⎪ ⎪ stage (μ (χ)) ⎪ ((0.7 − χ) /(0.7 − 0.6)), if χ ∈ 􏼂 0.6 0.7 􏼃 ⎪ DB C ⎪ ⎪ ⎩ ⎭ 0, otherwise ((χ − 0.6) /(0.7 − 0.6)), if χ ∈ 􏼂 0.6 0.7 􏼃 ⎧ ⎪ ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ 1, if χ ∈ 􏼂 0.7 0.8 􏼃 (μ (χ)) � , DBC−stage 4 ⎪ ⎪ ((0.9 − χ) /(0.9 − 0.8)), if χ ∈ 0.8 0.9 ⎪ 􏼂 􏼃 ⎪ ⎪ ⎪ ⎩ ⎭ 0, otherwise ((χ − 0.8)/(0.9 − 0.8)), if χ ∈ 􏼂 0.8 0.9 􏼃 ⎧ ⎪ ⎫ ⎪ ⎨ ⎬ (μ (χ)) � 1, if χ ∈ 􏼂 0.9 1 􏼃 DBC−stage 5 ⎪ ⎪ ⎩ ⎭ 0, otherwise Table 5: Lookup table for layer 4 for BCP-T1F expert system. Rules MRI CT PET Results 1 NT N1 SHB 2 NT N2 SHB Severe infection 3 NT N3 SHB 4 NT N1 BEN Stage 0 5 NT N2 BEN 6 LS N1 BEN Stage 1 7 LS N2 BEN 8 LS N2 SHB Stage 2 9 LS N3 SHB 10 HS N1 BEN Stage 3 11 HS N3 BEN 12 VHS N2 BEN 13 VHS N3 SHB Stage 4 14 HS N1 SHB 15 HS N3 SHB (μ (o)) is considered as node 2, (μ physical quantity into a measurement. Multiple sensors are CT nodes mode MRI−tumor size (])) is taken into account as low size, (μ (θ)) is connected in the form of topology with the sensor board. PET−benign spread in whole body, and (μ (χ)) turns out to be Each sensor node acquires a subset of the collected samples DBC−stage concluded as Stage 2. for locally compressing and summarizing from the random (μ (o)) is considered as node 1, (μ signal. CT nodes mode MRI−tumor size (])) is taken into account as no tumor, (μ (θ)) is PET−benign benign, and (μ (χ)) turns out to be concluded as Stage 0. DBC−stage 3.2.2. Preprocessing. Data preprocessing is a data mining (μ (o)) is considered as node 2, (μ CT nodes mode MRI−tumor size technique collecting the data from the patients which in- (])) is taken into account as low size, (μ (θ)) is be- PET−benign volves transforming raw data into an understandable format. nign, and (μ (χ)) turns out to be concluded as Stage 1. DBC−stage Real-world data is often incomplete, is inconsistent, and is likely to contain many errors. In this step, we handle the missing values using mean, mode, and so on. We also 3.2. SVM-Based System Model mitigate the noisy data using the moving average method in 3.2.1. Sensor Data. Heterogeneous sensors are collecting which we used five-filter size. Data preprocessing prepares continuously environmental data. It is transforming a raw data for further processing. CT (nodes) PET Ultrasound Biopsy gold (type) Journal of Healthcare Engineering 9 0.8 0.7 0.8 0.6 0.6 0.5 0.4 0.4 0.3 0.2 0.2 0.8 10 0.8 9 7 8 6 0.6 0.6 0.4 0.4 0.2 0.2 1 0 0 (a) (b) 0.7 0.8 0.6 0.6 0.5 0.4 0.4 0.2 0.3 11 0.2 8 15 7 15 15 6 5 5 (c) (d) Figure 3: (a) Rule surface for PETand MRI tumor size. (b) Rule surface for ultrasound and mammography. (c) Rule surface for CT (nodes) and MRI tumor size. (d) Rule surface for biopsy gold (type) and biopsy gold (severity). Figure 4: Lookup diagram for proposed BCP-T1F expert system. In this article, Figure 5 has proposed a new system through picturing of the proposed BCP-SVM system model for breast cancer control using support vector model. With the help of this model, we can witness that the machine system in ML [48] and for breast cancer pre- data gained from the Internet of medical things is utilized diction BCP-SVM. )is model depicts the whole process in sensory layer. )is fed data can be updated with the help Mammography Biopsy gold (severity) MRI tumor size (cm) MRI tumor size (cm) Diagnosis breast cancer Diagnosis breast cancer DBCT DBI-RADS 10 Journal of Healthcare Engineering sensors. )e layer named sensory layer has all the pa- j cos(θ) � , rameters which will be employed to predict cancer. )e ‖j‖ outcome generated is in the form of raw data. )e raw data (10) will be fed into the preprocessing layer. Data preprocessing j cos(c) � . prepares raw data for further processing. )e raw data goes ‖j‖ through the managing, moving, and normalization in the preprocessing layer. )e portable standard was employed Equation (3) can also be written as to eliminate inconsistencies from the data which is done in the previous layers termed as a preprocessing layer. After w � (cos(θ), cos(c)), the data from preprocessing layer jumps on to the appli- → → cation layer, this layer of various parameters which are used i · j � ‖i‖‖j‖cos(θ), in the application layer finds out the breast cancer ma- lignancy. )e layer is divided into two halves known as θ � η − c cos(θ) � cos(η − c) performance layer and prediction layer. In the prediction layer, Support Vector Machine is used to find out the breast � cos(η)cos(c) + sin(η)sin(c) cancer, and in the performance layer, it is used to evaluate the results produced by the prediction layer performance i j i j 1 1 2 2 � + which are shown in Figure 5. )e application layer eval- ‖i‖ ‖j‖ ‖i‖ ‖j‖ (11) uates the data being fed into this layer which then gives out whether the accuracy is achieved or not. i j + i j 1 1 2 2 � , )e proposed model is categorized into five different ‖j‖‖j‖ layers. If the trained accuracy is achieved, then it is passed onto the cloud for the further proceeding for the validation i j + i j 1 1 2 2 i · j � ‖i‖‖j‖􏼢 􏼣, process. Cloud stores the data whether it is for training ‖i‖‖j‖ process or for testing process. From cloud, data is being received for the validation process. )e trained data or input is fed into the cloud to i · j � 􏽘 i j . k k determine an evaluation system for the testing purposes. It is k�1 fed into the cloud and then forwarded to the preprocessing )e dot product can be computed as the above equation layer [51] where data is improved by handling missing values for n-dimensional vectors. and errors; then finally it is transferred to the further Let diagnosis. As we know, the equation of the line is z � m(i · j + y). (12) If sign (z) > 0, then it is correctly classified, and if sign (z) j � xj + y, (5) 2 1 < 0, then it is incorrectly classified. where “x” is slope of a line and “y” is the intersect; therefore, Given a dataset D, we compute f on a training dataset: xj − j + y � 0. (6) 1 2 z � m (i · j + y). (13) k k → → Let j � (j , j ) and i � (x, −1); then the above 1 2 )en Z which is called functional margin of the dataset is equation can be written as as follows: → → (7) i · j � 0. Z � min z . k (14) k�1.....n )is equation is obtained from 2-dimensional vectors. It Taking hyperplanes, the hyperplane with the largest Z also works for different number of dimensions; (6) depicts will be commendatory selected. )e geometric margin of the the general equation of hyperlane. → → dataset is denoted by Z. )e main goal is to take into account )e direction of a vector j � (j , j ) is written as i 1 2 an optimal hyperplane, which means finding the values of i and is defined as and y of the optimal hyperplane. j j 1 2 i � + , )e Lagrangian function shows the following equation: (8) ‖j‖ ‖j‖ where ψ(i, y, x) � i · i − 􏽘 c [m: (i · j + y) − 1], (15) 􏽱�������������� � 2 k�1 2 2 2 2 (9) ‖j‖ � j j j . . . j . 1 + 2 + 3 + n ϕ ψ(i, y, x) � i − 􏽘 c m j � 0, As we know, (16) i k k k k�1 Journal of Healthcare Engineering 11 Preprocessor Application layer Sensor layer layer Performance layer Prediction layer Handling Supervised Radius_mean missing machine value learning Texture mean Required Accuracy Yes learning Perimeter mean accuracy IoMT achived Raw Normali data zations No Support Miss rate vector machine Moving Concave point average Validation/Testing Radius_mean Texture mean Perimeter mean Yes Recommend Diagnose to a doctor Data Preprocessing for IoMT proposed model base No Discard Concave point Figure 5: Proposed BCP-SVM expert system methodology. n n n i(c, y) � 􏽘 c − 􏽘 􏽘 c c m m j j , (19) ϕ ψ(i, y, x) � − 􏽘 c m � 0. (17) k k k k l k l y k k k�1 k�1 l�1 k�1 From (16) and (17), we get and thus i � 􏽘 c m j , k k k n n k�1 max 􏽐 c − 􏽘 􏽘 c c m m j j i k l k l k l (18) k�1 k�1 l�1 􏽘 c m � 0. k k (20) k�1 Subject to c ≥ 0, k � 1, . . . , n, 􏽐 c m � 0. k k k After substituting the Lagrangian function ψ, we get k�1 Training 12 Journal of Healthcare Engineering Proposed BP-TF precision chart 120.00 96.56% 100.00 80.00 60.00 40.00 20.00 3.44% 0.00 Proposed BP-TF Precision Miss rate Figure 6: Precision chart of proposed BCP-T1F. )e expansion of the Lagrangian multipliers method to )e number of support vectors is S; we will have the the Karush–Kuhn–Tucker conditions can be done; the hyperplane. To make predictions, hyperplane is used. And constraints will bear disproportion. )e Kar- the hypothesis function is as follows: ush–Kuhn–Tucker commendatory conditions will be +1 if i · j + y≥ 0 expressed as h i 􏼁 � 􏼢 􏼣. (28) −1 if i · j + y< 0 c m i · j + y − 1 � 0, (21) 􏼂 􏼁 􏼃 k k k where j is the optimal point, c is the positive value, and α )e above point which arises on the hyperplane will be for the other points are ≈ 0. considered as class +1 (breast cancer found) and the point So, which lies down the hyperplane will be categorized as −1 (breast cancer not found). m i · j + y􏼁 − 1􏼁 � 0. (22) k k So, basically, the objective of the SM algorithm is to find )ese are called support vectors, which are the closest a hyperplane which could separate the data accurately and points to the hyperplane. According to (22), we need to find the best one, which is often referred to as the optimal hyperplane. i − 􏽘 c m j � 0, k k k k�1 4. Simulation and Results (23) i � 􏽘 c m j . k k k MATLAB 2019a is used for simulation purpose. Section 4.1 k�1 contains the results of proposed fuzzy-based model and Section 4.2 contains the result of proposed SVM-based To compute the value of y, we get model. m i · j + y􏼁 − 1􏼁 � 0. (24) k k Multiplying both sides by m in (24), then we get 4.1. Fuzzy Results. For the constructive results, MATLAB 2 ∗ R2019a is used as a tool so as to gather the stimulation of (25) m j · j + y􏼁 − m 􏼁 � 0, k k results taking algorithm development along with it; it also where m � 1; takes prototyping into account. )e interpretation of the results is being developed by taking the 12 total inputs and 4 i · j + y􏼁 − m 􏼁 � 0, k k (26) outputs variables for fuzzy logic. When layer 1 shows the y � m − i · j . k k symptoms to be found in a particular person, then it rushes to the second layer in which mammography and ultrasound )en are done to do the initial treatment so as to assure that something is fishy going on; in this research, the proposed y � 􏽘 m − i · j􏼁 . (27) k k BCP-T1F system not only diagnoses the disease but also k�1 shows the different levels. When jumping towards layer 3, it Journal of Healthcare Engineering 13 Table 6: Training of the proposed BCP-SVM system model during the prediction of breast cancer. Proposed BCP-SVM system model (70% of sample data in training) Total number of samples (N � 399) Result (output) (O , O ) B M Expected output (E , E ) O (breast cancer) positive O (benign) negative B M B M Input E � 250 positive 248 2 E � 149 negative 5 144 Table 7: Validation of the proposed BCP-SVM system model during the prediction of breast cancer. Proposed BCP-SVM system model (30% of sample data in validation) Total number of samples (N � 170) Result (output) (O , O ) B M Expected output (E , E ) O (breast cancer) positive O (benign) negative B M B M Input E � 107 positive 106 1 E � 63 negative 4 59 Table 8: Performance evaluation of proposed BCP-SVM system model in validation and training using different statistical measures. Miss False False Likelihood Positive Negative Likelihood Sensitivity Specificity Accuracy rate positive negative ratio prediction prediction ratio positive (%) value value negative value value (0.9863) (0.9802) (0.9825) (0.0198) (0.0137) (0.9664) (0.992) Training 1.75 49.81 0.0139 98.63% 98.02% 98.25% 1.98% 1.37% 96.64% 99.2% (0.9833) (0.9636) (0.9706) (0.0364) (0.0167) (0.9365) (0.9906) Validation 2.94 27.01 0.0173 98.33% 96.36% 97.06% 3.64% 1.67% 93.65% 99.06% Table 9: Comparison results of the proposed BCP-T1F and BCP-SVM system with literature. Training Literature Accuracy (%) Miss rate (%) ANN [19] 91.10 8.9 BCP ANN [8] 92.10 7.90 ANN [11] 94 6.0 ANN-ELM [8] 96.40 3.6 ANN [17] 91.1 8.9 Proposed BCP-T1F 96.56 3.44 Proposed BCP-SVM 97.06 2.94 depicts that the biopsy gold determines the type and severity Tables 6 and 7 conclude the training and validation with of the lump, whether it is cancerous or not. )en the third respect to precision rate and miss rate. SVM algorithm has layer comes; it involves PET, MRI, and CT which in turns been implemented to the dataset [15] of 569 sets of records; gives the size of the tumor and to which extent it is can- moreover, this data is divided into training constitutes of cerous. Figure 6 has clearly stated the precision of a proposed 70% (399 samples) and 30% (170 samples) for the men- system. )e proposed system BCP-T1F shows the precision tioned purposes training and validation. Various statistical rate of 96.56 percent and the miss rate of the BCP-T1F comes measures used for comparing as well as performance are out to be 3.44 percent. )is proposed system is providing the calculated with different metrics named sensitivity, spec- ificity, and accuracy, whereas the true positivity is accurate results for the corresponding type and severity level. expressed in sensitivity and accurate negative as specificity. )e following parameters are derived by the formulas given 4.2. SVM Results. )e simulation of MATLAB R2019a tool as follows: is employed to assume and predict the breast cancer. 14 Journal of Healthcare Engineering Precision chart 97.06 96.4 96.5 92.1 91.1 8.9 8.6 7.9 3.44 2.94 BPANN [8] ELM [8] ANN [11] ANN [19] Proposed Proposed BP-TF BB-SVM Precision Miss rate Figure 7: Comparisons with previous methods. benign and positive (1) value which shows the existence of O /E 􏼁 + O /E 􏼁 M B B M Misrate � , malignancy. E + E B M Table 6 shows the proposed BCP-SVM system model prediction of breast cancer during the training phase. Total O /E 􏼁 + O /E 􏼁 B B M M Precision � , 399 number of samples are used during training which is E + E B M further divided into 250,149 samples of positive and neg- ative, respectively. It is observed that 248 samples have O /E M M Sensitivity � , positive class which are correctly predicted and no breast O /E 􏼁 + O /E 􏼁 M M B B cancer (benign) is found but 02 records are incorrectly predicted as a negative which means breast cancer (malig- O /E 􏼁 B B Specificity � , nancy) is found. Similarly, total 149 samples are taken, O /E 􏼁 + O /E 􏼁 􏼁 M M B B wherein the case of negative means congestion is found, in which 144 samples are correctly predicted as a negative O /E 􏼁 B B which means breast cancer is found and 05 samples are Negative prediction value � , O /E + O /E 􏼁 􏼁 􏼁 M B B E invalidly predicted as a positive which means no breast cancer is found, while actually breast cancer exists there. O /E M M Table 7 shows the proposed BCP-SVM system model Positive prediction value � , O /E 􏼁 + O /E 􏼁 􏼁 M M B M prediction of breast cancer during validation phase. Total 170 numbers of samples are used during training which False positive ratio � 1 − specificity, further are divided into 107,63 samples of positive and negative, respectively. It is observed that 106 samples of False negative ratio � 1 − sensitivity, positive class have no breast cancer found and also are correctly predicted but 01 records are incorrectly predicted sensitivity as a negative which means breast cancer is found, while Likelihood ratio positive � , (1 − specificity) actually breast cancer does not exist. Similarly, total 63 samples are taken in the case of negative which means breast (1 − sensitivity) cancer is found, in which 59 samples are correctly predicted Likelihood ratio negative � . specificity as a negative which means breast cancer is found and 04 samples are invalidly predicted as a positive which means no (29) breast cancer is found, while actually breast cancer existed )e proposed BCP-SVM system model calculates the there. predicted output as negative (−1) and positive (1). )e re- Table 8 shows the proposed BCP-T1F-SVM system sultant output of value negative (−1) shows that there is model performance in terms of sensitivity, specificity, Journal of Healthcare Engineering 15 [2] A. Osareh and B. Shadgar, “Machine learning techniques to precision, and miss rate during training and testing phase. It diagnose breast cancer,” in Proceedings of the 2010 5th In- clearly shows that the proposed BCP-T1F-SVM system ternational Symposium on Health Informatics and Bio- during training gives 98.63%, 98.02%, 98.25%, and 1.75% informatics, pp. 114–120, IEEE, Antalya, Turkey, April 2010. sensitivity, specificity, accuracy, and miss rate, respectively. [3] D. R. Cox and D. Oakes, Analysis of Survival Data, Chapman And during testing, the proposed BCP-T1F-SVM system and Hall, London, UK, 1984. gives 98.33%, 96.36%, 97.06%, and 2.94% sensitivity, spec- [4] H. Brenner, O. Gefeller, and T. Hakulinen, “A computer ificity, accuracy, and miss rate, respectively. In addition, program for period analysis of cancer patient survival,” Eu- some more statistical measures are added to predict the ropean Journal of Cancer, vol. 38, no. 5, pp. 690–695, 2002. values such as false positive, false negative, likelihood ratio [5] I. Maglogiannis, E. Zafiropoulos, and I. Anagnostopoulos, negative, and positive and positive and negative prediction “An intelligent system for automated breast cancer diagnosis values give the result during training 1.98%, 1.37%, 0.0139, and prognosis using SVM based classifiers,” Applied Intelli- gence, vol. 30, no. 1, pp. 24–36, 2009. 49.81, 96.64%, and 99.2%. And during testing, the proposed [6] E. Gerasimova-Chechkina, B. Toner, Z. Marin et al., “Com- TCC-SVM system gives 3.64%, 1.67%, 0.0173, 27.01, 93.65%, bining multifractal analyses of digital mammograms and and 99.06%, respectively. infrared thermograms to assist in early breast cancer diag- Table 9 and Figure 7 show the performance of the nosis,” AIP Conference Proceedings, vol. 1760, no. 1, Article ID proposed BCP-T1F-SVM system model using fuzzy logic 020018, 2016. and SVM with previous approaches given in the literature [7] T. S. Subashini, V. Ramalingam, and S. Palanivel, “Breast mass BCP-ANN [8], ANN-ELM [8], and ANN [11, 17]. classification based on cytological patterns using RBFNN and SVM,” Expert Systems with Applications, vol. 36, no. 3, pp. 5284–5290, 2009. 5.Conclusion [8] C. Prasetyo, A. Kardiana, and R. Yuliwulandari, “Breast cancer diagnosis using artificial neural networks with extreme For the constructive results, MATLAB R2019a is used as a learning techniques,” International Journal of Advanced Re- tool so as to gather the stimulation of results taking algo- search in Artificial Intelligence, vol. 3, no. 7, pp. 10–14, 2014. rithm development along with it; it also takes prototyping [9] American Cancer Society, Breast Cancer Facts & Figures, into account. )e interpretation of the results is being de- American Cancer Society, Atlanta, GA, USA, 2007. veloped by taking the 12 total inputs and 4 outputs variables [10] R. L. Siegel, K. D. Miller, and A. Jemal, “Cancer statistics, for fuzzy logic and 30 inputs and 1 output variables for SVM. 2017,” CA: A Cancer Journal for Clinicians, vol. 67, no. 1, )e main goal is to analyze critically the different dimen- pp. 7–30, 2017. sions of breast cancer or any type of cancerous disease. )e [11] R. Setiono, “Extracting rules from pruned neural networks for proposed system BCP-T1F-SVM is to devise a type of expert breast cancer diagnosis,” Artificial Intelligence in Medicine, system to diagnose breast cancer and its stages. )e reports vol. 8, no. 1, pp. 37–51, 1996. [12] I. W. Tsang, J. T. Kwok, and P. M. Cheung, “Core vector on the basis of which the expert system analysis has been machines: fast SVM training on very large data sets,” Journal carried out were yielded by Cavan General Hospital Lis- of Machine Learning Research, vol. 6, pp. 363–392, 2005. daran, Cavan, Ireland. )is expert system can be employed [13] W. H. Wolberg and O. L. Mangasarian, “Multisurface method by the medical specialists and nonspecialist also. In this of pattern separation for medical diagnosis applied to breast article, the proposed principal expert system BCP-T1F cytology,” Proceedings of the National Academy of Sciences, achieved a precision of 96.56 percent, which in turn also vol. 87, no. 23, pp. 9193–9196, 1990. accords the 3.44 percent of miss rate. )e proposed BCP- [14] N. R. Goulding, J. D. Marquez, E. M. Prewett et al., “Ultra- SVM is proven to provide an accuracy of 97.06 percent sonic imaging techniques for breast cancer detection,” AIP which in turn also accords the 2.94 percent of miss rate. Both Conference Proceedings, vol. 975, no. 1, pp. 656–663, 2008. the proposed BCP-T1F and BCP-SVM systems give more [15] https://www.kaggle.com/uciml/breast-cancer-wisconsin-data. accuracy as compared to a previous published approach. [16] J. P. Suckling, “)e mammographic image analysis society digital mammogram database,” Exerpta Medica, vol. 1069, pp. 375–386, 1994. Data Availability [17] R. W. Brause, “Medical analysis and diagnosis by neural networks,” in Proceedings of the 2001 International Sympo- )e simulation data used to support the findings of this sium on Medical Data Analysis, pp. 1–13, Madrid, Spain, study are available from the corresponding author upon October 2001. request. [18] N. Tariq, “Breast cancer detection using artificial neural network,” Journal of Molecular Biomarkers and Diagnosis, vol. 9, no. 1, pp. 1–6, 2017. Conflicts of Interest [19] K. Crockett, Z. Bandar, D. Mclean, and J. O’Shea, “On con- structing a fuzzy inference framework using crisp decision trees,” )e authors declare that there are no conflicts of interest Fuzzy Sets and Systems, vol. 157, no. 21, pp. 2809–2832, 2006. regarding the publication of this article. [20] M. U. Khan, J. P. Choi, H. Shin, and M. Kim, “Predicting breast cancer survivability using fuzzy decision trees for References personalized healthcare,” in Proceedings of the 2008 30th Annual International Conference of the IEEE Engineering in [1] N. J. Bundred, “Prognostic and predictive factors in breast Medicine and Biology Society, pp. 5148–5151, IEEE, Van- cancer,” Cancer Treatment Reviews, vol. 27, no. 3, pp. 137–142, couver, Canada, August 2008. 2001. 16 Journal of Healthcare Engineering [21] L. G. Sison and E. K. Chong, “Fuzzy modeling by induction layered socio-fuzzy inference system,” Journal of Intelligent & and pruning of decision trees,” in Proceedings of 1994 9th Fuzzy Systems, vol. 37, no. 2, pp. 2769–2784, 2019. IEEE International Symposium on Intelligent Control, [38] A. Fatima, M. A. Khan, S. Abbas, M. Waqas, L. Anum, and M. Asif, “Evaluation of planet factors of smart city through pp. 166–171, IEEE, Columbus, OH, USA, August 1994. [22] M. Umano, H. Okamoto, I. Hatano et al., “Generation of fuzzy multi-layer fuzzy logic (MFL),” Ne ISC International Journal decision trees by fuzzy ID3 algorithm and its application to of Information Security, vol. 11, no. 3, pp. 51–58, 2019. diagnosis by gas in oil,” in Proceedings of the 1994 [39] S. Y. Siddiqui, S. A. Hussnain, A. H. Siddiqui et al., “Diagnosis Japan—USA Symposium on Flexible Automation, Ann Arbor, of arthritis using adaptive hierarchical Mamdani fuzzy type-1 MI, USA, 1994. expert system,” ICST Transactions on Scalable Information [23] H. Ian, Witten, and Eibe Frank, Data Mining: Practiacal Systems, Article ID 161439, 2019. Machine Learning Tools and Techniques, Morgan Kaufman, [40] A. Atta, S. Abbas, M. A. Khan, G. Ahmed, and U. Farooq, “An San Fransisco, CA, USA, 2nd edition, 2005. adaptive approach: smart traffic congestion control system,” [24] Nurhasanah, J. Sampurno, I. D. Faryuni, and O. Ivansyah, Journal of King Saud University-Computer and Information “Automated analysis of image mammogram for breast cancer Sciences, 2018. diagnosis,” AIP Conference Proceedings, vol. 1719, no. 1, [41] M. A. Khan, M. Umair, M. A. Saleem, M. N. Ali, and S. Abbas, Article ID 030036, 2016. “CDE using improved opposite based swarm optimization for MIMO systems,” Journal of Intelligent & Fuzzy Systems, [25] S. Wadhwani, T. Singh, and S. Singh Bhadauoria, “BCD- clustering algorithm for breast cancer diagnosis,” Interna- vol. 37, no. 1, pp. 687–692, 2019. tional Journal of Scientific and Research Publications, vol. 2, [42] M. A. Khan, M. Umair, and M. A. S. Choudhry, “GA based pp. 1–5, 2012. adaptive receiver for MC-CDMA system,” Turkish Journal of [26] V. Balanica, ˘ I. Dumitrache, M. Caramihai, W. Rae, and Electrical Engineering & Computer Sciences, vol. 23, no. 1, C. Herbst, “Evaluation of breast cancer risk by using fuzzy pp. 2267–2277, 2015. logic,” University Politehnica of Bucharest Scientific Bulletin, [43] M. A. Khan, M. Umair, and M. A. S. Choudry, “Island dif- Series C, University Politehnica of Bucharest, vol. 73, no. 1, ferential evolution based adaptive receiver for MC-CDMA pp. 53–64, Bucharest, Romania, 2011. system,” in Proceedings of the 2015 International Conference [27] D. B. Fogel, E. C. Wasson, E. M. Boughton, and V. W. Porto, on Information and Communication Technologies (ICICT), “Evolving artificial neural networks for screening features pp. 1–6, IEEE, Karachi, Pakistan, December 2015. from mammograms,” Artificial Intelligence in Medicine, [44] M. N. Ali, M. A. Khan, M. Adeel, and M. Amir, “Genetic vol. 14, no. 3, pp. 317–326, 1998. algorithm based adaptive receiver for MC-CDMA system with variation in mutation operator,” International Journal of [28] R. Bellazzi, L. Ironi, R. Guglielmann, and M. Stefanelli, “Qualitative models and fuzzy systems: an integrated ap- Computer Science and Information Security, vol. 14, no. 9, proach for learning from data,” Artificial Intelligence in p. 296, 2016. Medicine, vol. 14, no. 1-2, pp. 5–28, 1998. [45] M. Umair, M. A. Khan, and M. A. S. Choudry, “Island genetic [29] S. C. Bagui, S. Bagui, K. Pal, and N. R. Pal, “Breast cancer algorithm based MUD for MC-CDMA system,” in Proceed- detection using rank nearest neighbor classification rules,” ings of the 2015 International Conference on Information and Pattern Recognition, vol. 36, no. 1, pp. 25–34, 2003. Communication Technologies (ICICT), pp. 1–6, IEEE, Karachi, [30] R. Bellazzi, R. Guglielmann, and L. Ironi, “Qualitative models Pakistan, December 2015. and fuzzy systems: an integrated approach to system iden- [46] M. Umair, M. A. Khan, and M. A. S. Choudry, “GA backing to tification,” Soft Computing Applications, pp. 83–94, Springer, STBC based MC-CDMA systems,” in Proceedings of the 2013 Berlin, Germany, 2003. 4th International Conference on Intelligent Systems, Modelling and Simulation, pp. 503–506, IEEE, Bangkok, )ailand, [31] G. J. Miao, K. H. Miao, and J. H. Miao, “Neural pattern recognition model for breast cancer diagnosis,” Journal of January 2013. Selected Areas in Bioinformatics, pp. 1–8, 2012. [47] I. Kashif, A. K. Muhammad, A. Sagheer, H. Zahid, and [32] A. Torres and J. J. Nieto, “Fuzzy logic in medicine and bio- F. Areej, “Intelligent transportation system (ITS) for smart- informatics,” Journal of Biomedicine and Biotechnology, cities using Mamdani fuzzy inference system,” International vol. 2006, Article ID 91908, 7 pages, 2006. Journal of Advanced Computer Science and Applications [33] N. Cristianini, J. Kandola, A. Elisseeff, and J. Shawe-Taylor, (IJACSA), vol. 9, no. 2, pp. 94–105, 2018. “On optimizing Kernel alignment,” 2001, http://www.iipl. [48] A. M. Alqudah, H. M. S. Algharib, A. M. S. Algharib, and fudan.edu.cn/∼zhangjp/literatures/cluster%20analysis/01087. H. M. S. Algharib, “Computer aided diagnosis system for ps. automatic two stages classification of breast mass in digital [34] G. B. Huang and H. A. Babri, “Upper bounds on the number mammogram images,” Biomedical Engineering: Applications, of hidden neurons in feedforward networks with arbitrary Basis and Communications, vol. 31, no. 1, Article ID 1950007, bounded nonlinear activation functions,” IEEE Transactions 2019. [49] A. Alqudah and A. M. Alqudah, “Sliding window based on Neural Networks, vol. 9, no. 1, pp. 224–229, 1998. [35] A. K. Singh and B. Gupta, “A novel approach for breast cancer support vector machine system for classification of breast detection and segmentation in a mammogram,” Procedia cancer using histopathological microscopic images,” IETE Computer Science, vol. 54, pp. 676–682, 2015. Journal of Research, pp. 1–9, 2019. [36] J. Dheeba, N. Albert Singh, and S. Tamil Selvi, “Computer- [50] H. Wang, Z. Zhang, and T. Taleb, “Editorial: special issue on aided detection of breast cancer on mammograms: a swarm security and privacy of IoT,” World Wide Web, vol. 21, no. 1, intelligence optimized wavelet neural network approach,” pp. 1–6, 2018. Journal of Biomedical Informatics, vol. 49, pp. 45–52, 2014. [51] MathWorks, Inc, “MATLAB documentation,” 2017, https:// [37] S. Hussain, S. Abbas, T. Sohail, M. Adnan Khan, and A. Athar, www.mathworks.com/help/matlab/index.html?s_tid�gn_ “Estimating virtual trust of cognitive agents using multi loc_drop. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Healthcare Engineering Hindawi Publishing Corporation

Cloud-Based Breast Cancer Prediction Empowered with Soft Computing Approaches

Download PDF
16 pages

Loading next page...
 
/lp/hindawi-publishing-corporation/cloud-based-breast-cancer-prediction-empowered-with-soft-computing-13IxqZ1MRX

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher
Hindawi Publishing Corporation
ISSN
2040-2295
eISSN
2040-2309
DOI
10.1155/2020/8017496
Publisher site
See Article on Publisher Site

Abstract

Hindawi Journal of Healthcare Engineering Volume 2020, Article ID 8017496, 16 pages https://doi.org/10.1155/2020/8017496 Research Article Cloud-Based Breast Cancer Prediction Empowered with Soft Computing Approaches 1,2 1,3 1 4 Farrukh Khan, Muhammad Adnan Khan , Sagheer Abbas , Atifa Athar, 1,5 1,5 6 Shahan Yamin Siddiqui, Abdul Hannan Khan, Muhammad Anwaar Saeed, 1,2 and Muhammad Hussain Department of Computer Science, National College of Business Administration and Economics, Lahore, Pakistan Department of Computer Science, Lahore Institute of Science and Technology, Lahore, Pakistan Department of Computer Science, Lahore Garrison University, Lahore, Pakistan Department of Computer Science, CUI, Lahore Campus, Pakistan Department of Computer Science, Minhaj University, Lahore, Pakistan Department of Computer Science, Virtual University, Islamabad, Pakistan Correspondence should be addressed to Muhammad Adnan Khan; madnankhan@ncbae.edu.pk Received 18 January 2020; Accepted 30 April 2020; Published 19 May 2020 Academic Editor: Norio Iriguchi Copyright © 2020 Farrukh Khan et al. )is is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. )e developing countries are still starving for the betterment of health sector. )e disease commonly found among the women is breast cancer, and past researches have proven results that if the cancer is detected at a very early stage, the chances to overcome the disease are higher than the disease treated or detected at a later stage. )is article proposed cloud-based intelligent BCP-T1F- SVM with 2 variations/models like BCP-T1F and BCP-SVM. )e proposed BCP-T1F-SVM system has employed two main soft computing algorithms. )e proposed BCP-T1F-SVM expert system specifically defines the stage and the type of cancer a person is suffering from. Expert system will elaborate the grievous stages of the cancer, to which extent a patient has suffered. )e proposed BCP-SVM gives the higher precision of the proposed breast cancer detection model. In the limelight of breast cancer, the proposed BCP-T1F-SVM expert system gives out the higher precision rate. )e proposed BCP-T1F expert system is being employed in the diagnosis of breast cancer at an initial stage. Taking different stages of cancer into account, breast cancer is being dealt by BCP-T1F expert system. )e calculations and the evaluation done in this research have revealed that BCP-SVM is better than BCP-T1F. )e BCP-T1F concludes out the 96.56 percentage accuracy, whereas the BCP-SVM gives accuracy of 97.06 percentage. )e above unleashed research is wrapped up with the conclusion that BCP-SVM is better than the BCP-T1F. )e opinions have been recommended by the medical expertise of Sheikh Zayed Hospital Lahore, Pakistan, and Cavan General Hospital, Lisdaran, Cavan, Ireland. formed in the breast cells; when these cells begin to grow 1.Introduction irregularly in a human body, it results in flaking and redness Apparently, the diagnosis and the scrutiny of the breast of the breast. )e cancer is still considered as the undiag- cancer disease have always been a decisive and critical one in nosed and untreated disease in various parts of the world. the regard of medical department. )e cancerous lumps )e questions arise, why the cancer still has the strong roots form in a particular area of the body when the human cells in patients? Why still breast cancer remains undiagnosed begin to produce rapidly beyond the expected limit. )e amongst the women? Why the ratio of mortality among cancerous lumps which are also termed as tumors are women due to breast cancer is still constant? )e breast comprised of two kinds: one is the benign and the other is cancer should be diagnosed at an early stage so that the malignant. Breast cancer is considered as a lump that is condition does not persist and many lives could be saved. If 2 Journal of Healthcare Engineering it gets diagnosed initially, then the chances to overcome the 2. Literature Review disease are certainly higher. Women are being impinged by To avoid the faulty reasoning processes, errors, continuous this disease commonly. So if cancer remains undiagnosed, failures, lack of knowledge, and failures in rules of logics in then it may lead to death [1, 2]. )e risk factors evolve detection of tumors, researchers have started to advance new because of which breast cancer is induced may be the genetic methodologies, models, and tools. Bearing a clear aim, reasons, alcohol intake, dense tissues in breast, radiation various systems and proposed models have been put forward exposure, age, and so on. Since 1989, the survival rate in successful identification of breast cancer. Among women amongst masses has been immensely improved due to in America at the rate of one in three cancers approximately, modern technologies introduced in screening and treat- breast cancer is the most frequently diagnosed cancer among ment. )e recently conducted researches have shown that, in the women [6, 12, 13]. Surgical biopsies confirm malignancy 2017, 252,710 women were diagnosed with this disease; with high level of sensitivity but are considered costly and approximately 40,610 according to the statistics were more can affect patient’s psychology as well. )is research dem- likely to die from breast cancer. )e radical steps in reducing onstrates novel approach by using morphological operators the risk factors of this disease are the awareness of the in- and clustering algorithm fuzzy c-m to identify malignant duction causes of the breast cancer. )e early symptoms of lump in mammography automatically [6, 14–17]. In the cancer and screening can lessen the factors of cancer in- article, initial identification of tumor, fuzzy system’s various fringing masses [3]. applications, and algorithms have been proposed [18, 19]. In )e physical diagnosis of this mortal disease involves some previous studies on FDTs, proposed approaches focus the breast exam, imaging tests, biopsy, blood tests, and so on modification of decision tree pruning algorithm and on. )e initial blood marker tests which are CA 15.3, require fuzzy parameters to be set by domain experts. We TRU-QUANT, CA 125, CEA, and so on are done before opted to fuzzify already generated decision tree nodes to the treatment of this cancer. )e blood marker tests work relax the sharp decision boundaries. A similar kind of ap- as an initial indicator for this disease; in determining the proach is employed in [20, 21]. breast cancer, currently there are four main methods Fuzzy logic has been rarely used in cancer prognosis. which are employed in differentiating between the ma- Being noncrisp, it can act as a natural ally of a physician in lignant lump and benign lump: biopsy, fine needle as- prognostic decision-making process [22]. In the recent re- piration, mammography, and MRI [4, 5]. )e proposed searches, we have surveyed various types of research sce- methodology has employed the two algorithms to detect narios [21–23]; in the prognosis of cancer, the applications of an ailment efficiently. )e Mamdani inference system is various cases of machine learning techniques are contrib- used among the recent researches to detect a particular uting towards the advanced researches. Some of the basic kind of disease; this proposed research has compiled the trends which are encountered for the motivation of ex- previously done evaluations in comparison to this pro- periments include the following: the fuzzy logic has been posed methodology [6]. Mammography determines the used in the diagnosis of cancer rarely. Aiming for clear incomplete diagnosis of the infection; if the infection interpretation of a particular type of disease physicians, results conclude out to be negative, the infection turns out using “Black Box” models, approximately 70% of all re- as benign. Mammography determines whether there is searches reported making use of neural networks. )e any lump, no lump, or cancerous lump. Mammography majority of the manuscripts used machine learning tech- concludes the severity level and the type of cancer is niques independently without considering potential in the measured with the help of biopsy gold test. )e biopsy discussed manuscripts to cope up with each other in a hybrid gold test (type) will also determine the benign, ductal model. Lack of attention is paid to data size. Victor gives one carcinoma, invasive lobular carcinoma, inflammatory solution to computerized tool used for diagnosis of breast breast disease, and lobular carcinoma [7, 8]. cancer. )e fact is that fuzzy logic can substantially assist in )e recent researches show that, in 2017, 252,710 women diagnosis of breast cancer is being put forward in this paper were diagnosed with this disease; approximately 40,610 [24–26]. which were more likely to die from breast cancer [1, 9, 10]. Diagnosis of breast cancer is through fuzzy clustering )e physical diagnosis involves the breast exam, imaging with partial supervision [27]. ARTMAP approach gives tests, biopsy, blood tests, and so on. In regard of determining 97.2% accuracy, which represents the one-way approach in the breast cancer, currently there are four main methods the neural networks for the diagnosis of breast cancer [28]. which are used in differentiating between the malignant Classification exactness of over 95% was professed to be lump and benign lump: biopsy, fine needle aspiration, accomplished by utilizing little MIAS database; the proposed mammography, and MRI [2, 3, 11]. framework is being acknowledged for the findings of bosom In this research, fuzzy logic and SVM are used to find out malignancy dependent on outrageous learning machine the breast cancer by using different statistical measures such [29–32]. as accuracy, miss rate, specificity, sensitivity, false-positive Resisting a technique for classification of mammogram value, false-negative value, likelihood ratio positive, likeli- that is comprised of 4 phases, preprocessing stage utilized hood ratio negative, positive prediction value, and negative middle filter to upgrade nature of picture and to expel prediction values. With the help of these matrices, breast clamor from the picture. To check the variation from the cancer can be found more accurately as compared with the norm of the mammograms, ANN classifier was utilized to previous literature. Journal of Healthcare Engineering 3 group the picture into fitting class. Affectability, specificity, defined easily by the help of a medical expert. Type-1 fuzzy and precision asserted in the work were 72.72%, 93.6%, and logic rules are applied on inputs of fuzzy sets and then converted it into a fuzzy output. In this research, input 88.66% [18, 33]. A framework for the findings of bosom malignancy variables are used to propose a system to diagnose the dependent through feedforward networks was proposed. particular disease which is cardiac by using a fuzzy logic Prepreparing was done in two-phase foundation and second model. For the detection of cardiac disease, Support Vector was evacuating pectoral muscle. Hough change strategy was Machine (SVM) is a model that provides computational utilized for ROI. An aggregate of 32 dark dimensions and results which depend upon the structure and biological surface highlights is separated from mammograms. Preci- functions of neural networks. In the prediction layer, sion asserted by utilizing smaller than normal MIAS data- Support Vector Machine is used to find out the breast base was 94.06% [34]. cancer, and in the performance layer, it is used to evaluate A few systems have been sent to anticipate and perceive the results produced by the prediction layer performance. significant example for breast malignancy analysis. Data )e whole system process is shown in Figure 1, in which the data acquisition layer comprises the parameters of input. In mining further categorizes the different methods of the decision tree, ANN, RIPPERS classifier, and Support Vector this model, they will go for the neural system, where a Machine (SVM) to make a quick explanation and survey of trained algorithm is used to estimate breast cancer. At the the dataset regarding the cardiovascular disease. )e ex- industrial level, SVM is utilized and it gives accurate results. planation used the consideration and comparison of the SVM includes several neurons that are specifically orga- performance of the techniques which encompasses accuracy, nized. Neurons and influences among them are essential sensitivity, specificity, error rate, true positive rate, and false parts of an SVM. Neurons have handling features that co- positive rate [35, 36]. operate to overcome an issue. )is layer is used to examine Computational intelligence approaches like fuzzy system the breast cancer on the basis of thirty input parameters, [37–39], neural network [40], and swarm intelligence [41] which is termed as the scientific study of the models that are and evolutionary computing [42] like genetic algorithm statistical in nature and constitute algorithms which com- puter systems employ to perform a certain type of task with [43, 44], DE, Island GA [45], Island DE [46, 47], classifier [48], and SVM [49] are strong candidate solutions in the greater precision and as certainty. In the performance field of smart city [50], wireless communication, and so on. evaluation layer, precision and miss rate are determined. In the decisive area, the conclusion is made whether the breast 3.Proposed System Model Methodology cancer is identified or not. )e following methodology has been elaborated in Figure 1. First layer is the data acquisition layer which follows up with 3.1. Fuzzy System Methodology. Our proposed model breast the data collection of breast cancer. )e raw data attained cancer prediction (BCP), multilayered Mamdani fuzzy type- through the collection of breast cancer is then fed into the 1 inference system- (MFIS-) based expert system (BCP-T1F) preprocessing layer. )e preprocessing layer is a criterion to is explained in this section. )e BCP-T1F expert system consists of four layers as shown in Figure 2. In layer 1 named handle the missing values amongst the raw data; further- more, moving average and normalizations are being done in symptoms, the initial symptoms of breast cancer will be checked which are swelling, breast pain, redness, nipple the preprocessing layer. )e omissions and errors are being lessened through the standard portable. )en after the retraction, Family Inheritance Breast Cancer (FBIC), and completion of the previous layers, the preprocessing layer skin irritation. )is will find whether it is lump or cancer. then jumps on to the application layer. )e application layer If the system finds the symptoms in the patient, then is comprised of the prediction layer and the performance layer 2 diagnoses the breast cancer (no/yes) using two input evaluation layer. )e prediction layer specifically focuses on variables that are ultrasound and mammography. If the layer the two algorithms which are employed to determine the 2 diagnoses breast cancer, then layer 3 will be activated. Layer 3 predicts the type and severity of BCP based on two indispensable types of breast cancer through type-1 fuzzy logic and SVM just points out that something is fishy or not; input variables (biopsy gold severity) and (biopsy gold type). )en layer 4 will check the stage of cancer by three input that is, a person is suffering from the disease or not. )e two algorithms which aided the application layer are shown in variables that are MRI, CT, and PET which are shown in Figure 2. Figure 1. Type-1 fuzzy logic is an enabled system used to get accurate results from big data. )e performance evaluation Mathematically, the layers of the proposed BCP-T1F can layer calculates the accuracy and miss rate. Type-1 fuzzy be written as follows. logic constitutes of logical rules and these rules can be )is layer 1 can be written mathematically as μ � MFIS􏽨μ , μ , μ , μ , μ , μ 􏽩. (1) DBI,Layer1 swelling Breast pain Redness Skin irritation FBIC nipple−retraction 4 Journal of Healthcare Engineering Application layer Data Pre- Performance Prediction acquisition processing evaluation layer layer layer layer Handling missing Type-1 Malignancy Is breast value Refer to fuzzy cancer emergency logic Accuracy identified? Breast cancer Moving data average Raw collection data Miss rate Support Benign vector machine Normali- zations Discard Figure 1: Proposed intelligent breast cancer prediction model for BCP-T1F-SVM expert system. Biopsy gold type MRI Ultrasound Skin redness CT Biopsy gold type Mammography PET Lump/cancerous lump Mild-grad/High Breast pain severity/low severity Infected BCP Condition Condition Layer 3 Layer 4 Condition Redness Layer 1 Layer 2 No cancerous Normal No lump FBIC lump Discard Discard Discard Nipple retraction Figure 2: Proposed BCP-T1F expert system methodology. )e layer 2 can be written as )en layer 4 can be written as μ � MFIS μ , μ . (2) μ � MFIS􏼂μ , μ , μ 􏼃. (4) 􏽨 􏽩 DBC−ST, Layer 4 MRI CT PET DBI−RADS,Layer 2 Ultrasond Mammography )en layer 3 can be written as 3.1.1. Membership Functions. )e membership function of μ � MFIS􏽨μ , μ 􏽩. DBC−ST, Layer 3 Biopsy Gold−Type Biopsy Gold−Severity proposed BCP-T1F expert system yields the curve values (3) ranging between 0 and 1 and also dispenses a mathematical Journal of Healthcare Engineering 5 Table 1: Input and output variables membership functions used in the proposed BCP-T1F expert system. Sr. I/P parameters Mathematics of membership function number μ (α) � {max(min(1, ((0.5 − α)/0.1)), 0)}, Swelling−affected 1 Swelling (μ (α)) Swelling μ (α) � max(min(1, ((α − 0.4)/0.1) ), 0) { } Swelling−not affected μ (β) � 􏼈max(min(1, ((1.5 − β)/0.1)), 0)􏼉, Skin−irritation,affected 2 Skin irritation (μ (β)) Skin−irritation μ (β) � 􏼈max(min(1, ((β − 1.4)/0.1)), 0)􏼉 Skin−irritation,not affected μ (c) � 􏼈max(min(1, ((2.5 − c)/0.1)), 0)􏼉, Breast pain−affected 3 Breast pain (μ (c)) Breast−pain μ (c) � 􏼈max(min(1, ((c − 2.4)/0.1)), 0)􏼉 Breast pain−not affected μ (ρ) � max(min(1, ((3.5 − ρ)/0.1)), 0) , 􏼈 􏼉 Redness−affected 4 Redness (μ (ρ)) Redness μ (ρ) � 􏼈max(min(1, ((ρ − 3.4)/0.1)), 0)􏼉 Redness−not affected Family inheritance, 5 μ (ψ) � 􏼈max(min(1, ((4.5 − ψ)/0.1)), 0)􏼉, μ (ψ) � 􏼈max(min(1, ((ψ − 4.4)/0.1)), 0)􏼉 E,N E,P Breast cancer (μ (ψ)) FIBC Nipple retraction μ (φ) � 􏼈max(min(1, ((5.5 − φ)/0.1)), 0)􏼉, Nipple retraction−affected (μ (φ)) μ (φ) � max(min(1, ((φ − 5.4)/0.1)), 0) 􏼈 􏼉 Nipple−retraction Nipple retraction−not affected 1, if di € 􏼂 0 0.4 􏼃, ⎧ ⎪ ⎫ ⎪ ⎨ ⎬ μ (di) � ((0.5 − di)/0.1), if di€ 􏼂 0.4 0.5 􏼃, , DI−normal ⎪ ⎪ ⎩ ⎭ 0, otherwise 7 Diagnosis infection (μ (di)) DI ⎧ ⎪ ((di − 0.4)/0.1), if di€ 􏼂 0.4 0.5 􏼃 ⎫ ⎪ ⎨ ⎬ μ (di) � 1, if di€ 0.5 1 􏼂 􏼃 DI−infected ⎪ ⎪ ⎩ ⎭ 0, otherwise Table 2: Input and output variables membership functions used in the proposed BCP-T1F expert system. Sr. I/P parameters Mathematics of membership function number μ (λ) � {max(min(1, ((0.5 − λ)/0.1)), 0)}, Swelling−affected 1 Ultrasound (μ (λ)) Ultrasound μ (λ) � max(min(1, ((λ − 0.4)/0.1)), 0) { } Swelling−not affected μ (τ ) � 􏼈max(min(1, ((0.9 − τ )/0.1)), 0)􏼉, Mammography−incomplete 1 1 μ (τ ) � 􏼈max(min(((τ − 0.8)/0.1), 1, ((1.9 − τ )/0.1)))􏼉, Mammography−negative 2 2 2 μ (τ ) � 􏼈max(min(((τ − 1.8)/0.1), 1, ((2.9 − τ )/0.1)))􏼉, Mammography−benign 3 3 3 2 Mammography (μ (τ)) μ (τ ) � 􏼈max(min(((τ − 2.8)/0.1), 1, ((3.9 − τ )/0.1)))􏼉, Mammography Mammography−probably benign 4 4 4 μ (τ ) � max(min(((τ − 3.8)/0.1), 1, ((4.9 − τ )/0.1) )) , 􏼈 􏼉 Mammography−suspicious 5 5 5 μ (τ ) � 􏼈max(min(((τ − 4.8)/0.1), 1, ((5.9 − τ )/0.1)))􏼉, Mammography−suggested malignancy 6 6 6 μ (τ ) � max(min(1, ((τ − 5.8)/0.1), 0)) 􏼈 􏼉 Mammography−proven malignnancy 7 7 1, if rads€ 0 0.3 ⎧ ⎪ 􏼂 􏼃 ⎫ ⎪ ⎨ ⎬ μ (rads) � ((0.4 − rads)/0.1) if rads€ 􏼂 0.3 0.4 􏼃 , DBI−Rads−No lump ⎪ ⎪ ⎩ ⎭ 0, otherwise ((rads − 0.3)/0.1), if rads€ 􏼂 0.3 0.4 􏼃 ⎧ ⎪ ⎫ ⎪ ⎪ ⎪ Detection of breast imaging reporting ⎪ ⎪ ⎨ ⎬ 1, if rads€ 􏼂 0.4 0.6 􏼃 3 and database system score μ (rads) � , DBI−Rads−Lump ⎪ ⎪ ((0.7 − rads)/0.1) if rads 0.6 0.7 ⎪ 􏼂 􏼃 ⎪ ⎪ ⎪ (μ (rads)) ⎩ ⎭ DBI−rads 0, otherwise 1, if rads€ 􏼂 0.7 1 􏼃 ⎧ ⎪ ⎫ ⎪ ⎨ ⎬ μ (rads) � ((rads − 0.6)/0.1) if rads€ 􏼂 0.6 0.7 􏼃 DBI−Rads−Cancerous lump ⎪ ⎪ ⎩ ⎭ 0, otherwise form of the fuzzy logic that accords statistical values of both 3.1.2. Rules Table. )e proposed system BCP-T1F rules table the input and output variables. )e mathematical repre- usually relies on the expert system which constitutes of sixty- sentation of proposed BCP-T1F expert system yields four inputs and output rules for layer 1, fifteen output and member functions of layers 1–4 shown in Tables 1–4. )ese input rules for layer 2, eight input and output rules for layer membership functions are gathered after the consultation 3, and thirty I/O rules for layer 4. )is rule (Table 5) has been with the medical experts from Cavan General Hospital obtained with the assistance of the medical experts from Lisdaran, Cavan, Ireland. Cavan General Hospital Lisdaran, Cavan, Ireland. 6 Journal of Healthcare Engineering Table 3: Input and output variables membership functions used in the proposed BCP-T1F expert system. Sr. number I/P parameters Mathematics of membership function μ (δ ) � 􏼈max(min(1, ((5 − δ )/1)), 0)􏼉, Biopsy gold−severity−benign 1 1 Biopsy gold standard for severity 1 μ (δ ) � max(min(((δ − 4)/1), 1, ((10 − δ )/1)), 0) , 􏼈 􏼉 Biopsy gold−severity−cancer−severity low 2 2 2 (μ (δ)) Biopsy gold−severity μ (δ ) � 􏼈max(min(1, ((δ − 9)/1)), 0)􏼉 Biopsy gold−severity−cancer−severity high 3 3 μ (π ) � 􏼈max(min(1, ((5 − π )/1)), 0)􏼉, Biopsy gold−type−fine needle aspiration 1 1 Biopsy gold standard for type 2 μ (π ) � 􏼈max(min(((π − 4)/1), 1, ((10 − π )/1)), 0)􏼉, Biopsy gold−type−core needle biopsy 2 2 2 (μ (π)) Biopsy gold−type μ (π ) � 􏼈max(min(1, ((π − 9)/1)), 0)􏼉 Biopsy gold−type−surgical biopsy 3 3 1, if ∈ 􏼂 0 0.3 􏼃 ⎧ ⎪ ⎫ ⎪ ⎨ ⎬ μ (ρ ) � ((0.4 − ρ )/0.1), if ∈ 􏼂 0.3 0.4 􏼃 , DBCS−mild grade 1 1 ⎪ ⎪ ⎩ ⎭ 0, otherwise ((ρ − 0.3)/0.1), if ∈ 0.3 0.4 ⎧ ⎪ 􏼂 􏼃 ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ Diagnosis of breast cancer severity 1, if ∈ 􏼂 0.4 0.6 􏼃 3 μ (ρ ) � , DBCS−moderate severity 2 ⎪ ⎪ (μ (ρ)) ⎪ ((0.7 − ρ )/0.1), if ∈ 􏼂 0.6 0.7 􏼃 ⎪ DBCS 2 ⎪ ⎪ ⎩ ⎭ 0, otherwise ⎧ ⎪ 1, if rads ∈ 􏼂 0.7 1 􏼃 ⎫ ⎪ ⎨ ⎬ μ (ρ ) � ((ρ − 0.6)/0.1), if rads ∈ 0.6 0.7 􏼂 􏼃 DBCS−high severity 3 ⎪ 3 ⎪ ⎩ ⎭ 0, otherwise ⎧ ⎪ 1, if ∈ 􏼂 0 0.2 􏼃 ⎫ ⎪ ⎨ ⎬ μ (b ) � ((0.3 − b , )/0.1) if ∈ 􏼂 0.2 0.3 􏼃 DBCT−severity 1 1 ⎪ ⎪ ⎩ ⎭ 0, otherwise ((b − 0.2)/0.1), if ∈ 􏼂 0.2 0.3 􏼃 ⎧ ⎪ ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ 1, if ∈ 􏼂 0.3 0.4 􏼃 μ (b ) � , DBCT−ductal carcinoma 2 ⎪ ⎪ ⎪ ((0.5 − b )/0.1) if ∈ 􏼂 0.4 0.5 􏼃 ⎪ ⎪ ⎪ ⎩ ⎭ 0, otherwise 1, if ∈ 􏼂 0.5 0.6 􏼃 ⎧ ⎪ ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ Diagnosis of breast cancer ((b − 0.4)/0.1), if ∈ 􏼂 0.4 0.5 􏼃 4 μ (b ) � , DBCT−invasive lobular carcinoma 3 ⎪ ⎪ type (μ (b)) ⎪ ((0.7 − b )/0.1), if ∈ 􏼂 0.6 0.7 􏼃 ⎪ DBCT ⎪ 3 ⎪ ⎩ ⎭ 0, otherwise 1, if ∈ 0.6 0.7 ⎧ ⎪ 􏼂 􏼃 ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ ((b − 0.6)/0.1), if ∈ 􏼂 0.7 0.8 􏼃 μ (b ) � , DBCT−inflammatory breast disease 4 ⎪ ⎪ ⎪ ((0.9 − b )/0.1), if ∈ 􏼂 0.8 0.9 􏼃 ⎪ ⎪ ⎪ ⎩ ⎭ 0, otherwise ⎧ ⎪ 1, if ∈ 􏼂 0.9 1 􏼃 ⎫ ⎪ ⎨ ⎬ μ (b ) � ((b − 0.8)/0.1), if ∈ 0.8 0.9 􏼂 􏼃 DBCT−lobular carcinoma 5 ⎪ 5 ⎪ ⎩ ⎭ 0, otherwise 3.1.3. Rule Based. Rules are essential for input and output 3.1.6. Lookup Diagram. MATLAB R2019a tool is used for variables. Achievement of an adroit system is built based on demonstrating, imitation, algorithm expansion, prototyping, rules. Some of the rules are shown in Table 4. and various other fields. )is tool is well organized for software designing, data examination, conception, and calculations. For the simulation of results, three inputs and one output of BCP 3.1.4. Inference Engine. Inference engine is the most em- are used on layer 4 which are shown in Figure 4. phasized constituent of any decision-based expert system. In Figure 4 shows that (μ (o)) is considered as CT nodes mode this manuscript, BCP-T1F expert system has been employed node 2, (μ (])) is taken into account as low size, MRI−tumor size in layer 1, layer 2, layer 3, and layer 4. (μ (θ)) is spread in whole body, and (μ (χ)) turns PET DBC−stage out to be concluded as Stage 3. Similarly, Figure 4 also demonstrates the rule-based 3.1.5. Defuzzification. Defuzzification is the process of making knowledge; few of them are shown as follows: a measureable result in crusty logic, given fuzzy sets, and (μ (o)) is considered as node 3, CT nodes mode corresponding membership degrees. It is the process that plots a (μ (])) is taken into account as very high size, MRI−tumor size fuzzy set to a crisp set. It is characteristically needed in fuzzy (μ (θ)) is spread in whole body, and (μ (χ)) PET−benign DBC−stage control systems. In Figures 3(a)–3(d), the graphical illustrations turns out to be concluded as Stage 4. of defuzzifier of BCP-T1F expert system are obtainable. Journal of Healthcare Engineering 7 Table 4: Input and output variables membership functions used in the proposed BCP-T1F expert system. Sr. I/P parameters Mathematics of membership function number ⎧ ⎪ 1, if o ∈ 􏼂 0 7.9 􏼃 ⎫ ⎪ ⎨ ⎬ (μ (o)) � ((10 − o)/(10 − 9.9)), if o ∈ 9.9 10 , 􏼂 􏼃 CT nodes − mode 1 ⎪ ⎪ ⎩ ⎭ 0, otherwise ((o − 9.9) /(10 − 9.9)), if o ∈ 􏼂 9.9 10 􏼃 ⎧ ⎪ ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ Computed tomography (CT) 1, if o ∈ 􏼂 10 19.9 􏼃 1 (μ (o )) � , CT nodes − mode 1 2 ⎪ ⎪ (μ (o)) ⎪ ((20 − o) /(20 − 9.9)), if o ∈ 􏼂 19.9 20 􏼃 ⎪ CT nodes mode ⎪ ⎪ ⎩ ⎭ 0, otherwise ⎧ ⎪ ((o − 19.9)/(20 − 19.9)), if o ∈ 􏼂 19.9 20 􏼃 ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ 1, if o ∈ 􏼂 20 30 􏼃 (μ (o )) � CT nodes − mode 1 3 ⎪ ⎪ ⎪ ((20 − o) /(20 − 9.9)), if o ∈ 􏼂 19.9 20 􏼃 ⎪ ⎪ ⎪ ⎩ ⎭ 0, otherwise 1, if ] ∈ 0.02 0.1 ⎧ ⎪ 􏼂 􏼃 ⎫ ⎪ ⎨ ⎬ μ (]) � ((0.2 − ])/0.1) if ] ∈ 􏼂 0.1 0.2 􏼃 , MRI − tumor size− no ⎪ ⎪ ⎩ ⎭ 0, otherwise ((] − 0.1)/0.1), if ] ∈ 􏼂 0.1 0.2 􏼃 ⎧ ⎪ ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ 1, if ] ∈ 􏼂 0.2 1.9 􏼃 μ (]) � , MRI−tumor size−low ⎪ ⎪ ((2 − ])/0.1), if ] ∈ 1.9 2 ⎪ 􏼂 􏼃 ⎪ ⎪ ⎪ ⎩ ⎭ 0, otherwise 1, if ] ∈ 􏼂 2 4.9 􏼃 ⎧ ⎪ ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ Magnetic resonance imaging ((] − 1.9)/0.1), if ] ∈ 􏼂 1.9 2 􏼃 2 μ (]) � , MRI − tumor size − medium ⎪ ⎪ (μ (])) ⎪ ((5 − ])/0.1), if ] ∈ 􏼂 4.9 5 􏼃 ⎪ MRI−tumor size ⎪ ⎪ ⎩ ⎭ 0, otherwise ⎧ ⎪ 1, if ] ∈ 􏼂 5 7.9 􏼃 ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ ((] − 4.9)/0.1), if ] ∈ 􏼂 4.9 5 􏼃 μ (]) � , MRI − tumor size − high ⎪ ⎪ ⎪ ((8 − ])/0.1), if ] ∈ 􏼂 7.9 8 􏼃 ⎪ ⎪ ⎪ ⎩ ⎭ 0, otherwise ⎧ ⎪ 1, if ] ∈ 􏼂 8 11 􏼃 ⎫ ⎪ ⎨ ⎬ μ (]) � ((] − 7.9)/0.1) if ] ∈ 􏼂 7.9 8 􏼃 DBCT − lobular carcinoma ⎪ ⎪ ⎩ ⎭ 0, otherwise 1, if θ ∈ 0 0.4 ⎧ ⎪ 􏼂 􏼃 ⎫ ⎪ ⎨ ⎬ μ (θ) � ((0.5 − θ)/0.1) if θ ∈ 􏼂 0.4 0.5 􏼃 , PET − benign ⎪ ⎪ ⎩ ⎭ 0, otherwise Positron emission tomography (μ (θ)) PET ((θ − 0.4)/(0.5 − 0.4)), if θ ∈ 􏼂 0.4 0.5 􏼃 ⎧ ⎪ ⎫ ⎪ ⎨ ⎬ μ (θ) � 1, if θ ∈ 􏼂 0.5 1 􏼃 PET _in_whole_body ⎪ ⎪ spread ⎩ ⎭ 0, otherwise 8 Journal of Healthcare Engineering Table 4: Continued. Sr. I/P parameters Mathematics of membership function number ⎧ ⎪ 1, if χ ∈ 􏼂 0 0.2 􏼃 ⎫ ⎪ ⎨ ⎬ (μ (χ)) � , ((0.3 − χ)/(0.3 − 0.2)), if χ ∈ 􏼂 0.2 0.3 􏼃 DBC−stage 1 ⎪ ⎪ ⎩ ⎭ 0, otherwise ((χ − 0.2) ( / 0.3 − 0.2)), if χ ∈ 􏼂 0.2 0.3 􏼃 ⎧ ⎪ ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ 1, if χ ∈ 􏼂 0.3 0.4 􏼃 (μ (χ)) � , DBC−stage 2 ⎪ ⎪ ⎪ ((0.5 − χ) /(0.5 − 0.4)), if χ ∈ 􏼂 0.4 0.5 􏼃 ⎪ ⎪ ⎪ ⎩ ⎭ 0, otherwise ⎧ ⎪ ((χ − 0.4) /(0.5 − 0.4)), if χ ∈ 􏼂 0.4 0.5 􏼃 ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ Diagnosis of breast cancer 1, if χ ∈ 􏼂 0.5 0.6 􏼃 4 (μ (χ)) � , DBC−stage 3 ⎪ ⎪ stage (μ (χ)) ⎪ ((0.7 − χ) /(0.7 − 0.6)), if χ ∈ 􏼂 0.6 0.7 􏼃 ⎪ DB C ⎪ ⎪ ⎩ ⎭ 0, otherwise ((χ − 0.6) /(0.7 − 0.6)), if χ ∈ 􏼂 0.6 0.7 􏼃 ⎧ ⎪ ⎫ ⎪ ⎪ ⎪ ⎪ ⎪ ⎨ ⎬ 1, if χ ∈ 􏼂 0.7 0.8 􏼃 (μ (χ)) � , DBC−stage 4 ⎪ ⎪ ((0.9 − χ) /(0.9 − 0.8)), if χ ∈ 0.8 0.9 ⎪ 􏼂 􏼃 ⎪ ⎪ ⎪ ⎩ ⎭ 0, otherwise ((χ − 0.8)/(0.9 − 0.8)), if χ ∈ 􏼂 0.8 0.9 􏼃 ⎧ ⎪ ⎫ ⎪ ⎨ ⎬ (μ (χ)) � 1, if χ ∈ 􏼂 0.9 1 􏼃 DBC−stage 5 ⎪ ⎪ ⎩ ⎭ 0, otherwise Table 5: Lookup table for layer 4 for BCP-T1F expert system. Rules MRI CT PET Results 1 NT N1 SHB 2 NT N2 SHB Severe infection 3 NT N3 SHB 4 NT N1 BEN Stage 0 5 NT N2 BEN 6 LS N1 BEN Stage 1 7 LS N2 BEN 8 LS N2 SHB Stage 2 9 LS N3 SHB 10 HS N1 BEN Stage 3 11 HS N3 BEN 12 VHS N2 BEN 13 VHS N3 SHB Stage 4 14 HS N1 SHB 15 HS N3 SHB (μ (o)) is considered as node 2, (μ physical quantity into a measurement. Multiple sensors are CT nodes mode MRI−tumor size (])) is taken into account as low size, (μ (θ)) is connected in the form of topology with the sensor board. PET−benign spread in whole body, and (μ (χ)) turns out to be Each sensor node acquires a subset of the collected samples DBC−stage concluded as Stage 2. for locally compressing and summarizing from the random (μ (o)) is considered as node 1, (μ signal. CT nodes mode MRI−tumor size (])) is taken into account as no tumor, (μ (θ)) is PET−benign benign, and (μ (χ)) turns out to be concluded as Stage 0. DBC−stage 3.2.2. Preprocessing. Data preprocessing is a data mining (μ (o)) is considered as node 2, (μ CT nodes mode MRI−tumor size technique collecting the data from the patients which in- (])) is taken into account as low size, (μ (θ)) is be- PET−benign volves transforming raw data into an understandable format. nign, and (μ (χ)) turns out to be concluded as Stage 1. DBC−stage Real-world data is often incomplete, is inconsistent, and is likely to contain many errors. In this step, we handle the missing values using mean, mode, and so on. We also 3.2. SVM-Based System Model mitigate the noisy data using the moving average method in 3.2.1. Sensor Data. Heterogeneous sensors are collecting which we used five-filter size. Data preprocessing prepares continuously environmental data. It is transforming a raw data for further processing. CT (nodes) PET Ultrasound Biopsy gold (type) Journal of Healthcare Engineering 9 0.8 0.7 0.8 0.6 0.6 0.5 0.4 0.4 0.3 0.2 0.2 0.8 10 0.8 9 7 8 6 0.6 0.6 0.4 0.4 0.2 0.2 1 0 0 (a) (b) 0.7 0.8 0.6 0.6 0.5 0.4 0.4 0.2 0.3 11 0.2 8 15 7 15 15 6 5 5 (c) (d) Figure 3: (a) Rule surface for PETand MRI tumor size. (b) Rule surface for ultrasound and mammography. (c) Rule surface for CT (nodes) and MRI tumor size. (d) Rule surface for biopsy gold (type) and biopsy gold (severity). Figure 4: Lookup diagram for proposed BCP-T1F expert system. In this article, Figure 5 has proposed a new system through picturing of the proposed BCP-SVM system model for breast cancer control using support vector model. With the help of this model, we can witness that the machine system in ML [48] and for breast cancer pre- data gained from the Internet of medical things is utilized diction BCP-SVM. )is model depicts the whole process in sensory layer. )is fed data can be updated with the help Mammography Biopsy gold (severity) MRI tumor size (cm) MRI tumor size (cm) Diagnosis breast cancer Diagnosis breast cancer DBCT DBI-RADS 10 Journal of Healthcare Engineering sensors. )e layer named sensory layer has all the pa- j cos(θ) � , rameters which will be employed to predict cancer. )e ‖j‖ outcome generated is in the form of raw data. )e raw data (10) will be fed into the preprocessing layer. Data preprocessing j cos(c) � . prepares raw data for further processing. )e raw data goes ‖j‖ through the managing, moving, and normalization in the preprocessing layer. )e portable standard was employed Equation (3) can also be written as to eliminate inconsistencies from the data which is done in the previous layers termed as a preprocessing layer. After w � (cos(θ), cos(c)), the data from preprocessing layer jumps on to the appli- → → cation layer, this layer of various parameters which are used i · j � ‖i‖‖j‖cos(θ), in the application layer finds out the breast cancer ma- lignancy. )e layer is divided into two halves known as θ � η − c cos(θ) � cos(η − c) performance layer and prediction layer. In the prediction layer, Support Vector Machine is used to find out the breast � cos(η)cos(c) + sin(η)sin(c) cancer, and in the performance layer, it is used to evaluate the results produced by the prediction layer performance i j i j 1 1 2 2 � + which are shown in Figure 5. )e application layer eval- ‖i‖ ‖j‖ ‖i‖ ‖j‖ (11) uates the data being fed into this layer which then gives out whether the accuracy is achieved or not. i j + i j 1 1 2 2 � , )e proposed model is categorized into five different ‖j‖‖j‖ layers. If the trained accuracy is achieved, then it is passed onto the cloud for the further proceeding for the validation i j + i j 1 1 2 2 i · j � ‖i‖‖j‖􏼢 􏼣, process. Cloud stores the data whether it is for training ‖i‖‖j‖ process or for testing process. From cloud, data is being received for the validation process. )e trained data or input is fed into the cloud to i · j � 􏽘 i j . k k determine an evaluation system for the testing purposes. It is k�1 fed into the cloud and then forwarded to the preprocessing )e dot product can be computed as the above equation layer [51] where data is improved by handling missing values for n-dimensional vectors. and errors; then finally it is transferred to the further Let diagnosis. As we know, the equation of the line is z � m(i · j + y). (12) If sign (z) > 0, then it is correctly classified, and if sign (z) j � xj + y, (5) 2 1 < 0, then it is incorrectly classified. where “x” is slope of a line and “y” is the intersect; therefore, Given a dataset D, we compute f on a training dataset: xj − j + y � 0. (6) 1 2 z � m (i · j + y). (13) k k → → Let j � (j , j ) and i � (x, −1); then the above 1 2 )en Z which is called functional margin of the dataset is equation can be written as as follows: → → (7) i · j � 0. Z � min z . k (14) k�1.....n )is equation is obtained from 2-dimensional vectors. It Taking hyperplanes, the hyperplane with the largest Z also works for different number of dimensions; (6) depicts will be commendatory selected. )e geometric margin of the the general equation of hyperlane. → → dataset is denoted by Z. )e main goal is to take into account )e direction of a vector j � (j , j ) is written as i 1 2 an optimal hyperplane, which means finding the values of i and is defined as and y of the optimal hyperplane. j j 1 2 i � + , )e Lagrangian function shows the following equation: (8) ‖j‖ ‖j‖ where ψ(i, y, x) � i · i − 􏽘 c [m: (i · j + y) − 1], (15) 􏽱�������������� � 2 k�1 2 2 2 2 (9) ‖j‖ � j j j . . . j . 1 + 2 + 3 + n ϕ ψ(i, y, x) � i − 􏽘 c m j � 0, As we know, (16) i k k k k�1 Journal of Healthcare Engineering 11 Preprocessor Application layer Sensor layer layer Performance layer Prediction layer Handling Supervised Radius_mean missing machine value learning Texture mean Required Accuracy Yes learning Perimeter mean accuracy IoMT achived Raw Normali data zations No Support Miss rate vector machine Moving Concave point average Validation/Testing Radius_mean Texture mean Perimeter mean Yes Recommend Diagnose to a doctor Data Preprocessing for IoMT proposed model base No Discard Concave point Figure 5: Proposed BCP-SVM expert system methodology. n n n i(c, y) � 􏽘 c − 􏽘 􏽘 c c m m j j , (19) ϕ ψ(i, y, x) � − 􏽘 c m � 0. (17) k k k k l k l y k k k�1 k�1 l�1 k�1 From (16) and (17), we get and thus i � 􏽘 c m j , k k k n n k�1 max 􏽐 c − 􏽘 􏽘 c c m m j j i k l k l k l (18) k�1 k�1 l�1 􏽘 c m � 0. k k (20) k�1 Subject to c ≥ 0, k � 1, . . . , n, 􏽐 c m � 0. k k k After substituting the Lagrangian function ψ, we get k�1 Training 12 Journal of Healthcare Engineering Proposed BP-TF precision chart 120.00 96.56% 100.00 80.00 60.00 40.00 20.00 3.44% 0.00 Proposed BP-TF Precision Miss rate Figure 6: Precision chart of proposed BCP-T1F. )e expansion of the Lagrangian multipliers method to )e number of support vectors is S; we will have the the Karush–Kuhn–Tucker conditions can be done; the hyperplane. To make predictions, hyperplane is used. And constraints will bear disproportion. )e Kar- the hypothesis function is as follows: ush–Kuhn–Tucker commendatory conditions will be +1 if i · j + y≥ 0 expressed as h i 􏼁 � 􏼢 􏼣. (28) −1 if i · j + y< 0 c m i · j + y − 1 � 0, (21) 􏼂 􏼁 􏼃 k k k where j is the optimal point, c is the positive value, and α )e above point which arises on the hyperplane will be for the other points are ≈ 0. considered as class +1 (breast cancer found) and the point So, which lies down the hyperplane will be categorized as −1 (breast cancer not found). m i · j + y􏼁 − 1􏼁 � 0. (22) k k So, basically, the objective of the SM algorithm is to find )ese are called support vectors, which are the closest a hyperplane which could separate the data accurately and points to the hyperplane. According to (22), we need to find the best one, which is often referred to as the optimal hyperplane. i − 􏽘 c m j � 0, k k k k�1 4. Simulation and Results (23) i � 􏽘 c m j . k k k MATLAB 2019a is used for simulation purpose. Section 4.1 k�1 contains the results of proposed fuzzy-based model and Section 4.2 contains the result of proposed SVM-based To compute the value of y, we get model. m i · j + y􏼁 − 1􏼁 � 0. (24) k k Multiplying both sides by m in (24), then we get 4.1. Fuzzy Results. For the constructive results, MATLAB 2 ∗ R2019a is used as a tool so as to gather the stimulation of (25) m j · j + y􏼁 − m 􏼁 � 0, k k results taking algorithm development along with it; it also where m � 1; takes prototyping into account. )e interpretation of the results is being developed by taking the 12 total inputs and 4 i · j + y􏼁 − m 􏼁 � 0, k k (26) outputs variables for fuzzy logic. When layer 1 shows the y � m − i · j . k k symptoms to be found in a particular person, then it rushes to the second layer in which mammography and ultrasound )en are done to do the initial treatment so as to assure that something is fishy going on; in this research, the proposed y � 􏽘 m − i · j􏼁 . (27) k k BCP-T1F system not only diagnoses the disease but also k�1 shows the different levels. When jumping towards layer 3, it Journal of Healthcare Engineering 13 Table 6: Training of the proposed BCP-SVM system model during the prediction of breast cancer. Proposed BCP-SVM system model (70% of sample data in training) Total number of samples (N � 399) Result (output) (O , O ) B M Expected output (E , E ) O (breast cancer) positive O (benign) negative B M B M Input E � 250 positive 248 2 E � 149 negative 5 144 Table 7: Validation of the proposed BCP-SVM system model during the prediction of breast cancer. Proposed BCP-SVM system model (30% of sample data in validation) Total number of samples (N � 170) Result (output) (O , O ) B M Expected output (E , E ) O (breast cancer) positive O (benign) negative B M B M Input E � 107 positive 106 1 E � 63 negative 4 59 Table 8: Performance evaluation of proposed BCP-SVM system model in validation and training using different statistical measures. Miss False False Likelihood Positive Negative Likelihood Sensitivity Specificity Accuracy rate positive negative ratio prediction prediction ratio positive (%) value value negative value value (0.9863) (0.9802) (0.9825) (0.0198) (0.0137) (0.9664) (0.992) Training 1.75 49.81 0.0139 98.63% 98.02% 98.25% 1.98% 1.37% 96.64% 99.2% (0.9833) (0.9636) (0.9706) (0.0364) (0.0167) (0.9365) (0.9906) Validation 2.94 27.01 0.0173 98.33% 96.36% 97.06% 3.64% 1.67% 93.65% 99.06% Table 9: Comparison results of the proposed BCP-T1F and BCP-SVM system with literature. Training Literature Accuracy (%) Miss rate (%) ANN [19] 91.10 8.9 BCP ANN [8] 92.10 7.90 ANN [11] 94 6.0 ANN-ELM [8] 96.40 3.6 ANN [17] 91.1 8.9 Proposed BCP-T1F 96.56 3.44 Proposed BCP-SVM 97.06 2.94 depicts that the biopsy gold determines the type and severity Tables 6 and 7 conclude the training and validation with of the lump, whether it is cancerous or not. )en the third respect to precision rate and miss rate. SVM algorithm has layer comes; it involves PET, MRI, and CT which in turns been implemented to the dataset [15] of 569 sets of records; gives the size of the tumor and to which extent it is can- moreover, this data is divided into training constitutes of cerous. Figure 6 has clearly stated the precision of a proposed 70% (399 samples) and 30% (170 samples) for the men- system. )e proposed system BCP-T1F shows the precision tioned purposes training and validation. Various statistical rate of 96.56 percent and the miss rate of the BCP-T1F comes measures used for comparing as well as performance are out to be 3.44 percent. )is proposed system is providing the calculated with different metrics named sensitivity, spec- ificity, and accuracy, whereas the true positivity is accurate results for the corresponding type and severity level. expressed in sensitivity and accurate negative as specificity. )e following parameters are derived by the formulas given 4.2. SVM Results. )e simulation of MATLAB R2019a tool as follows: is employed to assume and predict the breast cancer. 14 Journal of Healthcare Engineering Precision chart 97.06 96.4 96.5 92.1 91.1 8.9 8.6 7.9 3.44 2.94 BPANN [8] ELM [8] ANN [11] ANN [19] Proposed Proposed BP-TF BB-SVM Precision Miss rate Figure 7: Comparisons with previous methods. benign and positive (1) value which shows the existence of O /E 􏼁 + O /E 􏼁 M B B M Misrate � , malignancy. E + E B M Table 6 shows the proposed BCP-SVM system model prediction of breast cancer during the training phase. Total O /E 􏼁 + O /E 􏼁 B B M M Precision � , 399 number of samples are used during training which is E + E B M further divided into 250,149 samples of positive and neg- ative, respectively. It is observed that 248 samples have O /E M M Sensitivity � , positive class which are correctly predicted and no breast O /E 􏼁 + O /E 􏼁 M M B B cancer (benign) is found but 02 records are incorrectly predicted as a negative which means breast cancer (malig- O /E 􏼁 B B Specificity � , nancy) is found. Similarly, total 149 samples are taken, O /E 􏼁 + O /E 􏼁 􏼁 M M B B wherein the case of negative means congestion is found, in which 144 samples are correctly predicted as a negative O /E 􏼁 B B which means breast cancer is found and 05 samples are Negative prediction value � , O /E + O /E 􏼁 􏼁 􏼁 M B B E invalidly predicted as a positive which means no breast cancer is found, while actually breast cancer exists there. O /E M M Table 7 shows the proposed BCP-SVM system model Positive prediction value � , O /E 􏼁 + O /E 􏼁 􏼁 M M B M prediction of breast cancer during validation phase. Total 170 numbers of samples are used during training which False positive ratio � 1 − specificity, further are divided into 107,63 samples of positive and negative, respectively. It is observed that 106 samples of False negative ratio � 1 − sensitivity, positive class have no breast cancer found and also are correctly predicted but 01 records are incorrectly predicted sensitivity as a negative which means breast cancer is found, while Likelihood ratio positive � , (1 − specificity) actually breast cancer does not exist. Similarly, total 63 samples are taken in the case of negative which means breast (1 − sensitivity) cancer is found, in which 59 samples are correctly predicted Likelihood ratio negative � . specificity as a negative which means breast cancer is found and 04 samples are invalidly predicted as a positive which means no (29) breast cancer is found, while actually breast cancer existed )e proposed BCP-SVM system model calculates the there. predicted output as negative (−1) and positive (1). )e re- Table 8 shows the proposed BCP-T1F-SVM system sultant output of value negative (−1) shows that there is model performance in terms of sensitivity, specificity, Journal of Healthcare Engineering 15 [2] A. Osareh and B. Shadgar, “Machine learning techniques to precision, and miss rate during training and testing phase. It diagnose breast cancer,” in Proceedings of the 2010 5th In- clearly shows that the proposed BCP-T1F-SVM system ternational Symposium on Health Informatics and Bio- during training gives 98.63%, 98.02%, 98.25%, and 1.75% informatics, pp. 114–120, IEEE, Antalya, Turkey, April 2010. sensitivity, specificity, accuracy, and miss rate, respectively. [3] D. R. Cox and D. Oakes, Analysis of Survival Data, Chapman And during testing, the proposed BCP-T1F-SVM system and Hall, London, UK, 1984. gives 98.33%, 96.36%, 97.06%, and 2.94% sensitivity, spec- [4] H. Brenner, O. Gefeller, and T. Hakulinen, “A computer ificity, accuracy, and miss rate, respectively. In addition, program for period analysis of cancer patient survival,” Eu- some more statistical measures are added to predict the ropean Journal of Cancer, vol. 38, no. 5, pp. 690–695, 2002. values such as false positive, false negative, likelihood ratio [5] I. Maglogiannis, E. Zafiropoulos, and I. Anagnostopoulos, negative, and positive and positive and negative prediction “An intelligent system for automated breast cancer diagnosis values give the result during training 1.98%, 1.37%, 0.0139, and prognosis using SVM based classifiers,” Applied Intelli- gence, vol. 30, no. 1, pp. 24–36, 2009. 49.81, 96.64%, and 99.2%. And during testing, the proposed [6] E. Gerasimova-Chechkina, B. Toner, Z. Marin et al., “Com- TCC-SVM system gives 3.64%, 1.67%, 0.0173, 27.01, 93.65%, bining multifractal analyses of digital mammograms and and 99.06%, respectively. infrared thermograms to assist in early breast cancer diag- Table 9 and Figure 7 show the performance of the nosis,” AIP Conference Proceedings, vol. 1760, no. 1, Article ID proposed BCP-T1F-SVM system model using fuzzy logic 020018, 2016. and SVM with previous approaches given in the literature [7] T. S. Subashini, V. Ramalingam, and S. Palanivel, “Breast mass BCP-ANN [8], ANN-ELM [8], and ANN [11, 17]. classification based on cytological patterns using RBFNN and SVM,” Expert Systems with Applications, vol. 36, no. 3, pp. 5284–5290, 2009. 5.Conclusion [8] C. Prasetyo, A. Kardiana, and R. Yuliwulandari, “Breast cancer diagnosis using artificial neural networks with extreme For the constructive results, MATLAB R2019a is used as a learning techniques,” International Journal of Advanced Re- tool so as to gather the stimulation of results taking algo- search in Artificial Intelligence, vol. 3, no. 7, pp. 10–14, 2014. rithm development along with it; it also takes prototyping [9] American Cancer Society, Breast Cancer Facts & Figures, into account. )e interpretation of the results is being de- American Cancer Society, Atlanta, GA, USA, 2007. veloped by taking the 12 total inputs and 4 outputs variables [10] R. L. Siegel, K. D. Miller, and A. Jemal, “Cancer statistics, for fuzzy logic and 30 inputs and 1 output variables for SVM. 2017,” CA: A Cancer Journal for Clinicians, vol. 67, no. 1, )e main goal is to analyze critically the different dimen- pp. 7–30, 2017. sions of breast cancer or any type of cancerous disease. )e [11] R. Setiono, “Extracting rules from pruned neural networks for proposed system BCP-T1F-SVM is to devise a type of expert breast cancer diagnosis,” Artificial Intelligence in Medicine, system to diagnose breast cancer and its stages. )e reports vol. 8, no. 1, pp. 37–51, 1996. [12] I. W. Tsang, J. T. Kwok, and P. M. Cheung, “Core vector on the basis of which the expert system analysis has been machines: fast SVM training on very large data sets,” Journal carried out were yielded by Cavan General Hospital Lis- of Machine Learning Research, vol. 6, pp. 363–392, 2005. daran, Cavan, Ireland. )is expert system can be employed [13] W. H. Wolberg and O. L. Mangasarian, “Multisurface method by the medical specialists and nonspecialist also. In this of pattern separation for medical diagnosis applied to breast article, the proposed principal expert system BCP-T1F cytology,” Proceedings of the National Academy of Sciences, achieved a precision of 96.56 percent, which in turn also vol. 87, no. 23, pp. 9193–9196, 1990. accords the 3.44 percent of miss rate. )e proposed BCP- [14] N. R. Goulding, J. D. Marquez, E. M. Prewett et al., “Ultra- SVM is proven to provide an accuracy of 97.06 percent sonic imaging techniques for breast cancer detection,” AIP which in turn also accords the 2.94 percent of miss rate. Both Conference Proceedings, vol. 975, no. 1, pp. 656–663, 2008. the proposed BCP-T1F and BCP-SVM systems give more [15] https://www.kaggle.com/uciml/breast-cancer-wisconsin-data. accuracy as compared to a previous published approach. [16] J. P. Suckling, “)e mammographic image analysis society digital mammogram database,” Exerpta Medica, vol. 1069, pp. 375–386, 1994. Data Availability [17] R. W. Brause, “Medical analysis and diagnosis by neural networks,” in Proceedings of the 2001 International Sympo- )e simulation data used to support the findings of this sium on Medical Data Analysis, pp. 1–13, Madrid, Spain, study are available from the corresponding author upon October 2001. request. [18] N. Tariq, “Breast cancer detection using artificial neural network,” Journal of Molecular Biomarkers and Diagnosis, vol. 9, no. 1, pp. 1–6, 2017. Conflicts of Interest [19] K. Crockett, Z. Bandar, D. Mclean, and J. O’Shea, “On con- structing a fuzzy inference framework using crisp decision trees,” )e authors declare that there are no conflicts of interest Fuzzy Sets and Systems, vol. 157, no. 21, pp. 2809–2832, 2006. regarding the publication of this article. [20] M. U. Khan, J. P. Choi, H. Shin, and M. Kim, “Predicting breast cancer survivability using fuzzy decision trees for References personalized healthcare,” in Proceedings of the 2008 30th Annual International Conference of the IEEE Engineering in [1] N. J. Bundred, “Prognostic and predictive factors in breast Medicine and Biology Society, pp. 5148–5151, IEEE, Van- cancer,” Cancer Treatment Reviews, vol. 27, no. 3, pp. 137–142, couver, Canada, August 2008. 2001. 16 Journal of Healthcare Engineering [21] L. G. Sison and E. K. Chong, “Fuzzy modeling by induction layered socio-fuzzy inference system,” Journal of Intelligent & and pruning of decision trees,” in Proceedings of 1994 9th Fuzzy Systems, vol. 37, no. 2, pp. 2769–2784, 2019. IEEE International Symposium on Intelligent Control, [38] A. Fatima, M. A. Khan, S. Abbas, M. Waqas, L. Anum, and M. Asif, “Evaluation of planet factors of smart city through pp. 166–171, IEEE, Columbus, OH, USA, August 1994. [22] M. Umano, H. Okamoto, I. Hatano et al., “Generation of fuzzy multi-layer fuzzy logic (MFL),” Ne ISC International Journal decision trees by fuzzy ID3 algorithm and its application to of Information Security, vol. 11, no. 3, pp. 51–58, 2019. diagnosis by gas in oil,” in Proceedings of the 1994 [39] S. Y. Siddiqui, S. A. Hussnain, A. H. Siddiqui et al., “Diagnosis Japan—USA Symposium on Flexible Automation, Ann Arbor, of arthritis using adaptive hierarchical Mamdani fuzzy type-1 MI, USA, 1994. expert system,” ICST Transactions on Scalable Information [23] H. Ian, Witten, and Eibe Frank, Data Mining: Practiacal Systems, Article ID 161439, 2019. Machine Learning Tools and Techniques, Morgan Kaufman, [40] A. Atta, S. Abbas, M. A. Khan, G. Ahmed, and U. Farooq, “An San Fransisco, CA, USA, 2nd edition, 2005. adaptive approach: smart traffic congestion control system,” [24] Nurhasanah, J. Sampurno, I. D. Faryuni, and O. Ivansyah, Journal of King Saud University-Computer and Information “Automated analysis of image mammogram for breast cancer Sciences, 2018. diagnosis,” AIP Conference Proceedings, vol. 1719, no. 1, [41] M. A. Khan, M. Umair, M. A. Saleem, M. N. Ali, and S. Abbas, Article ID 030036, 2016. “CDE using improved opposite based swarm optimization for MIMO systems,” Journal of Intelligent & Fuzzy Systems, [25] S. Wadhwani, T. Singh, and S. Singh Bhadauoria, “BCD- clustering algorithm for breast cancer diagnosis,” Interna- vol. 37, no. 1, pp. 687–692, 2019. tional Journal of Scientific and Research Publications, vol. 2, [42] M. A. Khan, M. Umair, and M. A. S. Choudhry, “GA based pp. 1–5, 2012. adaptive receiver for MC-CDMA system,” Turkish Journal of [26] V. Balanica, ˘ I. Dumitrache, M. Caramihai, W. Rae, and Electrical Engineering & Computer Sciences, vol. 23, no. 1, C. Herbst, “Evaluation of breast cancer risk by using fuzzy pp. 2267–2277, 2015. logic,” University Politehnica of Bucharest Scientific Bulletin, [43] M. A. Khan, M. Umair, and M. A. S. Choudry, “Island dif- Series C, University Politehnica of Bucharest, vol. 73, no. 1, ferential evolution based adaptive receiver for MC-CDMA pp. 53–64, Bucharest, Romania, 2011. system,” in Proceedings of the 2015 International Conference [27] D. B. Fogel, E. C. Wasson, E. M. Boughton, and V. W. Porto, on Information and Communication Technologies (ICICT), “Evolving artificial neural networks for screening features pp. 1–6, IEEE, Karachi, Pakistan, December 2015. from mammograms,” Artificial Intelligence in Medicine, [44] M. N. Ali, M. A. Khan, M. Adeel, and M. Amir, “Genetic vol. 14, no. 3, pp. 317–326, 1998. algorithm based adaptive receiver for MC-CDMA system with variation in mutation operator,” International Journal of [28] R. Bellazzi, L. Ironi, R. Guglielmann, and M. Stefanelli, “Qualitative models and fuzzy systems: an integrated ap- Computer Science and Information Security, vol. 14, no. 9, proach for learning from data,” Artificial Intelligence in p. 296, 2016. Medicine, vol. 14, no. 1-2, pp. 5–28, 1998. [45] M. Umair, M. A. Khan, and M. A. S. Choudry, “Island genetic [29] S. C. Bagui, S. Bagui, K. Pal, and N. R. Pal, “Breast cancer algorithm based MUD for MC-CDMA system,” in Proceed- detection using rank nearest neighbor classification rules,” ings of the 2015 International Conference on Information and Pattern Recognition, vol. 36, no. 1, pp. 25–34, 2003. Communication Technologies (ICICT), pp. 1–6, IEEE, Karachi, [30] R. Bellazzi, R. Guglielmann, and L. Ironi, “Qualitative models Pakistan, December 2015. and fuzzy systems: an integrated approach to system iden- [46] M. Umair, M. A. Khan, and M. A. S. Choudry, “GA backing to tification,” Soft Computing Applications, pp. 83–94, Springer, STBC based MC-CDMA systems,” in Proceedings of the 2013 Berlin, Germany, 2003. 4th International Conference on Intelligent Systems, Modelling and Simulation, pp. 503–506, IEEE, Bangkok, )ailand, [31] G. J. Miao, K. H. Miao, and J. H. Miao, “Neural pattern recognition model for breast cancer diagnosis,” Journal of January 2013. Selected Areas in Bioinformatics, pp. 1–8, 2012. [47] I. Kashif, A. K. Muhammad, A. Sagheer, H. Zahid, and [32] A. Torres and J. J. Nieto, “Fuzzy logic in medicine and bio- F. Areej, “Intelligent transportation system (ITS) for smart- informatics,” Journal of Biomedicine and Biotechnology, cities using Mamdani fuzzy inference system,” International vol. 2006, Article ID 91908, 7 pages, 2006. Journal of Advanced Computer Science and Applications [33] N. Cristianini, J. Kandola, A. Elisseeff, and J. Shawe-Taylor, (IJACSA), vol. 9, no. 2, pp. 94–105, 2018. “On optimizing Kernel alignment,” 2001, http://www.iipl. [48] A. M. Alqudah, H. M. S. Algharib, A. M. S. Algharib, and fudan.edu.cn/∼zhangjp/literatures/cluster%20analysis/01087. H. M. S. Algharib, “Computer aided diagnosis system for ps. automatic two stages classification of breast mass in digital [34] G. B. Huang and H. A. Babri, “Upper bounds on the number mammogram images,” Biomedical Engineering: Applications, of hidden neurons in feedforward networks with arbitrary Basis and Communications, vol. 31, no. 1, Article ID 1950007, bounded nonlinear activation functions,” IEEE Transactions 2019. [49] A. Alqudah and A. M. Alqudah, “Sliding window based on Neural Networks, vol. 9, no. 1, pp. 224–229, 1998. [35] A. K. Singh and B. Gupta, “A novel approach for breast cancer support vector machine system for classification of breast detection and segmentation in a mammogram,” Procedia cancer using histopathological microscopic images,” IETE Computer Science, vol. 54, pp. 676–682, 2015. Journal of Research, pp. 1–9, 2019. [36] J. Dheeba, N. Albert Singh, and S. Tamil Selvi, “Computer- [50] H. Wang, Z. Zhang, and T. Taleb, “Editorial: special issue on aided detection of breast cancer on mammograms: a swarm security and privacy of IoT,” World Wide Web, vol. 21, no. 1, intelligence optimized wavelet neural network approach,” pp. 1–6, 2018. Journal of Biomedical Informatics, vol. 49, pp. 45–52, 2014. [51] MathWorks, Inc, “MATLAB documentation,” 2017, https:// [37] S. Hussain, S. Abbas, T. Sohail, M. Adnan Khan, and A. Athar, www.mathworks.com/help/matlab/index.html?s_tid�gn_ “Estimating virtual trust of cognitive agents using multi loc_drop.

Journal

Journal of Healthcare Engineering – Hindawi Publishing Corporation

Published: May 19, 2020

There are no references for this article.