{"id":{"repo_id":"auckland-tech","oai_identifier":"oai:openrepository.aut.ac.nz:10292/13001"},"canonical_url":"https://search.dev.ndltd.org/etd/auckland-tech/oai:openrepository.aut.ac.nz:10292/13001","repository":{"repo_id":"auckland-tech","name":"AUT University","base_url":"https://openrepository.aut.ac.nz/server/oai/request"},"display":{"title":"Incorporating Service Proximity Into Web Service Recommendation Via Tensors Decomposition","abstract":"The sparseness of Mashup-API rating matrix coupled with cold-start and scalability issues have been identified as the most critical challenges that affect most Collaborative filtering based Web APIs recommendation solution. Sparseness deteriorates the rating prediction accuracy. Several Web-API recommendation approaches employ basic collaborative filtering technique which operates on second-order matrices or tensors by decomposing the Mashup-API interaction matrix into two low-rank matrix approximations, and then make prediction based on the factorized tensors. While most existing CF, Matrix factorization-based Web-API recommendation approaches have shown promising improvement in recommendation results, one limitation is that they only focuson2-dimensionaldatamodelinwhichhistoricalinteractionbetweenMashup-API are mainly used. However, recent works in recommendation domain show that by incorporating additional information into the rating data, Web-API rating prediction accuracy can be enhanced. Inspired by these works, this research proposes a collaborative Filtering method based Tensors factorization, an extension of Matrix factorization that exploits the ternary relation among three key entities in Web service ecosystem Mashup-API-Proximity. Modelling the Web-API rating data with Tensor decomposition technique enables incorporation proximity information as third additional entity into Web service recommendation application to improve prediction accuracy. Specifically, we employ High Order Singular Value Decomposition approach with regularization term to extend the traditional Mashup-API matrix into Mashup-API-Proximity tensors. 5 Experimental analysis on Programmable Web dataset shows promising results compare with some state-of-the-art approaches.","abstract_html":"The sparseness of Mashup-API rating matrix coupled with cold-start and scalability issues have been identified as the most critical challenges that affect most Collaborative filtering based Web APIs recommendation solution. Sparseness deteriorates the rating prediction accuracy. Several Web-API recommendation approaches employ basic collaborative filtering technique which operates on second-order matrices or tensors by decomposing the Mashup-API interaction matrix into two low-rank matrix approximations, and then make prediction based on the factorized tensors. While most existing CF, Matrix factorization-based Web-API recommendation approaches have shown promising improvement in recommendation results, one limitation is that they only focuson2-dimensionaldatamodelinwhichhistoricalinteractionbetweenMashup-API are mainly used. However, recent works in recommendation domain show that by incorporating additional information into the rating data, Web-API rating prediction accuracy can be enhanced. Inspired by these works, this research proposes a collaborative Filtering method based Tensors factorization, an extension of Matrix factorization that exploits the ternary relation among three key entities in Web service ecosystem Mashup-API-Proximity. Modelling the Web-API rating data with Tensor decomposition technique enables incorporation proximity information as third additional entity into Web service recommendation application to improve prediction accuracy. Specifically, we employ High Order Singular Value Decomposition approach with regularization term to extend the traditional Mashup-API matrix into Mashup-API-Proximity tensors. 5 Experimental analysis on Programmable Web dataset shows promising results compare with some state-of-the-art approaches.","abstract_has_math":false,"creators":["Wu, Zhentao"],"institution":null,"degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":[],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2019,"date_issued":"2019","date_published":"2019","updated_at":"2026-07-27T18:46:16Z","subjects":["Tensors Factorization","Programmable Web","High Order Singular Value Decomposition","Recommendation","Mashups","Web-API recommendation"],"languages":[],"rights":[],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier","label":"Identifier","values":["hdl:10292/13001"],"render_values":[{"text":"hdl:10292/13001","href":null,"code":true}]}]},"links":{"outbound_url":null,"outbound_label":null,"outbound_source":null},"metadata_groups":[{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.issued","label":"Date","values":["2019"]},{"key":"dc:type","label":"Dc Type","values":["Thesis"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Tensors Factorization","Programmable Web","High Order Singular Value Decomposition","Recommendation","Mashups","Web-API recommendation"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["hdl:10292/13001"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.other","label":"Dc Description Other","values":["The sparseness of Mashup-API rating matrix coupled with cold-start and scalability issues have been identified as the most critical challenges that affect most Collaborative filtering based Web APIs recommendation solution. Sparseness deteriorates the rating prediction accuracy. Several Web-API recommendation approaches employ basic collaborative filtering technique which operates on second-order matrices or tensors by decomposing the Mashup-API interaction matrix into two low-rank matrix approximations, and then make prediction based on the factorized tensors. While most existing CF, Matrix factorization-based Web-API recommendation approaches have shown promising improvement in recommendation results, one limitation is that they only focuson2-dimensionaldatamodelinwhichhistoricalinteractionbetweenMashup-API are mainly used. However, recent works in recommendation domain show that by incorporating additional information into the rating data, Web-API rating prediction accuracy can be enhanced. Inspired by these works, this research proposes a collaborative Filtering method based Tensors factorization, an extension of Matrix factorization that exploits the ternary relation among three key entities in Web service ecosystem Mashup-API-Proximity. Modelling the Web-API rating data with Tensor decomposition technique enables incorporation proximity information as third additional entity into Web service recommendation application to improve prediction accuracy. Specifically, we employ High Order Singular Value Decomposition approach with regularization term to extend the traditional Mashup-API matrix into Mashup-API-Proximity tensors. 5 Experimental analysis on Programmable Web dataset shows promising results compare with some state-of-the-art approaches."]},{"key":"dc:title","label":"Title","values":["Incorporating Service Proximity Into Web Service Recommendation Via Tensors Decomposition"]}]}],"canonical_facts":{"dc:date.issued":["2019"],"dc:description.other":["The sparseness of Mashup-API rating matrix coupled with cold-start and scalability issues have been identified as the most critical challenges that affect most Collaborative filtering based Web APIs recommendation solution. Sparseness deteriorates the rating prediction accuracy. Several Web-API recommendation approaches employ basic collaborative filtering technique which operates on second-order matrices or tensors by decomposing the Mashup-API interaction matrix into two low-rank matrix approximations, and then make prediction based on the factorized tensors. While most existing CF, Matrix factorization-based Web-API recommendation approaches have shown promising improvement in recommendation results, one limitation is that they only focuson2-dimensionaldatamodelinwhichhistoricalinteractionbetweenMashup-API are mainly used. However, recent works in recommendation domain show that by incorporating additional information into the rating data, Web-API rating prediction accuracy can be enhanced. Inspired by these works, this research proposes a collaborative Filtering method based Tensors factorization, an extension of Matrix factorization that exploits the ternary relation among three key entities in Web service ecosystem Mashup-API-Proximity. Modelling the Web-API rating data with Tensor decomposition technique enables incorporation proximity information as third additional entity into Web service recommendation application to improve prediction accuracy. Specifically, we employ High Order Singular Value Decomposition approach with regularization term to extend the traditional Mashup-API matrix into Mashup-API-Proximity tensors. 5 Experimental analysis on Programmable Web dataset shows promising results compare with some state-of-the-art approaches."],"dc:identifier":["hdl:10292/13001"],"dc:subject":["Tensors Factorization","Programmable Web","High Order Singular Value Decomposition","Recommendation","Mashups","Web-API recommendation"],"dc:title":["Incorporating Service Proximity Into Web Service Recommendation Via Tensors Decomposition"],"dc:type":["Thesis"]},"updated_at":"2026-07-27T18:46:16Z"}