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

Learn More →

Cross-domain based Event Recommendation using Tensor Factorization

Cross-domain based Event Recommendation using Tensor Factorization AbstractContext in the form of meta-data has been accreditedas an important component in cross-domain collaborativefiltering (CDCF). In this research paper CDCFconcept is used to exploit event information (context)from two UI matrices to allow the recommendation performanceof one domain (Facebook- User-Event Matrix)to benefit from the information from another domain(Bookmyshow- Event-Tag Matrix). The model based collaborativefiltering approach Tensor Factorization(TF) hasbeen used to integrate Facebook provided User-Event contextinformation with Bookmyshow Event-Tag context informationto recommend events. In contrast to the standardcollaborative tag recommendation, our CDCF approachuses one User-Event matrix of Facebook that takesanother Bookmyshow Event-Tag matrix as additional informant.The proposed cross-domain based Event Recommendationapproach is divided into three modules- i) datacollection which extracts the unstructured dataset fromthe two domains Bookmyshow and social networking siteFacebook using API’s; ii) data mapping module which isbasically used to integrate the common knowledge/ datathat can be shared between considered different domains(Facebook & Bookmyshow). This module integrates andreduces the data into structured events’ instances. As thedataset was collected from two different sites, an intersectionof both was taken out. Therefore this module is carefullydesigned according to reliability of information thatis common between two domains; iii) 3 order tensor factorizationand Latent Dirichlet Allocation (LDA) used formost preferable recommendation by less pertinent result reduction. The proposed 3 order tensor factorization is designedfor maximizing the mutual benefit from both theconsidered domains (organizer and user). Therefore providingthree recommendations: For organizers: 1) systemrecommends places to conduct specific event according tomaximum of attendees of a particular type of event at aspecific location; 2) recommending target audience to organizer:those who are interested to attend event on thebasis of past data for promotion purposes. For users: 3) recommendingevents to users of their interest on the basis ofpast record. Our result shows significant improvement inreduction of less relevant data and result effectiveness ismeasured through recall and precision. Reduction of lessrelevant recommendation is 64%, 72% and 63% for placerecommendation to organizer, target audience recommendationto organizer and event recommendation to userrespectively. The proposed tensor factorization approachachieved 68% precision, 15.5% recall in recommending attendeesto organizer and 62% precision, 13.4% recall forevent recommendation to user. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Open Computer Science de Gruyter

Cross-domain based Event Recommendation using Tensor Factorization

Download PDF
12 pages

Loading next page...
 
/lp/de-gruyter/cross-domain-based-event-recommendation-using-tensor-factorization-LVgKoWRL7i

References

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

Publisher
de Gruyter
Copyright
© 2016 A. Arora et al.
eISSN
2299-1093
DOI
10.1515/comp-2016-0011
Publisher site
See Article on Publisher Site

Abstract

AbstractContext in the form of meta-data has been accreditedas an important component in cross-domain collaborativefiltering (CDCF). In this research paper CDCFconcept is used to exploit event information (context)from two UI matrices to allow the recommendation performanceof one domain (Facebook- User-Event Matrix)to benefit from the information from another domain(Bookmyshow- Event-Tag Matrix). The model based collaborativefiltering approach Tensor Factorization(TF) hasbeen used to integrate Facebook provided User-Event contextinformation with Bookmyshow Event-Tag context informationto recommend events. In contrast to the standardcollaborative tag recommendation, our CDCF approachuses one User-Event matrix of Facebook that takesanother Bookmyshow Event-Tag matrix as additional informant.The proposed cross-domain based Event Recommendationapproach is divided into three modules- i) datacollection which extracts the unstructured dataset fromthe two domains Bookmyshow and social networking siteFacebook using API’s; ii) data mapping module which isbasically used to integrate the common knowledge/ datathat can be shared between considered different domains(Facebook & Bookmyshow). This module integrates andreduces the data into structured events’ instances. As thedataset was collected from two different sites, an intersectionof both was taken out. Therefore this module is carefullydesigned according to reliability of information thatis common between two domains; iii) 3 order tensor factorizationand Latent Dirichlet Allocation (LDA) used formost preferable recommendation by less pertinent result reduction. The proposed 3 order tensor factorization is designedfor maximizing the mutual benefit from both theconsidered domains (organizer and user). Therefore providingthree recommendations: For organizers: 1) systemrecommends places to conduct specific event according tomaximum of attendees of a particular type of event at aspecific location; 2) recommending target audience to organizer:those who are interested to attend event on thebasis of past data for promotion purposes. For users: 3) recommendingevents to users of their interest on the basis ofpast record. Our result shows significant improvement inreduction of less relevant data and result effectiveness ismeasured through recall and precision. Reduction of lessrelevant recommendation is 64%, 72% and 63% for placerecommendation to organizer, target audience recommendationto organizer and event recommendation to userrespectively. The proposed tensor factorization approachachieved 68% precision, 15.5% recall in recommending attendeesto organizer and 62% precision, 13.4% recall forevent recommendation to user.

Journal

Open Computer Science – de Gruyter

Published: Jan 1, 2016

There are no references for this article.