{"id":92,"date":"2020-12-03T09:19:53","date_gmt":"2020-12-03T09:19:53","guid":{"rendered":"http:\/\/dmslab.hkg03.bdysite.com\/?page_id=92"},"modified":"2021-12-07T03:04:19","modified_gmt":"2021-12-07T03:04:19","slug":"extremely-sparse-data","status":"publish","type":"page","link":"http:\/\/www.dmslab.cn\/?page_id=92","title":{"rendered":"Extremely Sparse Data"},"content":{"rendered":"\n<h2 class=\"wp-block-heading\"><strong>Extremely Sparse Data with Blocks-Coupled Non-negative Matrix Factorization<\/strong><\/h2>\n\n\n\n<p>Recommender systems have been comprehensively analyzed in the past \ndecade and made great achievement in various fields. Generally speaking,\n the recommendation of information of interests is based on the \npotential connections among users and items implied in \u2018User-Item \nMatrix\u2019. However, the exiting algorithm for recommendation will be \ndegraded and ever fail in the case of sparseness of matrix. To resolve \nthis problem, a new algorithm called B-NMF (Blocks-Coupled Non-negative \nMatrix Factorization) is proposed in this paper. With this algorithm: \n(1) the reconstruction performance of matrix of extreme sparseness is \nimproved as a result of blocking the matrix and modeling based on full \nuse of the coupling between blocks; (2) the coupling between different \nblocks is ensured via a coupling mechanism that imposes constraints on \nconsistency as the matrix is decomposed. In addition, we provide an \napproach to exploiting homophily effect in prediction via homophily \nregularization and thus, the coupling between blocks is improved via \nextra homophily regularization constraints.<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"648\" src=\"http:\/\/dmslab.hkg03.bdysite.com\/wp-content\/uploads\/2021\/12\/d52aa61f0a965936f9097bc4811ad9d3-sz_219359-1024x648.jpeg\" alt=\"\" class=\"wp-image-1608\" srcset=\"http:\/\/www.dmslab.cn\/wp-content\/uploads\/2021\/12\/d52aa61f0a965936f9097bc4811ad9d3-sz_219359-1024x648.jpeg 1024w, http:\/\/www.dmslab.cn\/wp-content\/uploads\/2021\/12\/d52aa61f0a965936f9097bc4811ad9d3-sz_219359-300x190.jpeg 300w, http:\/\/www.dmslab.cn\/wp-content\/uploads\/2021\/12\/d52aa61f0a965936f9097bc4811ad9d3-sz_219359-768x486.jpeg 768w, http:\/\/www.dmslab.cn\/wp-content\/uploads\/2021\/12\/d52aa61f0a965936f9097bc4811ad9d3-sz_219359.jpeg 1036w\" sizes=\"(max-width: 767px) 89vw, (max-width: 1000px) 54vw, (max-width: 1071px) 543px, 580px\" \/><\/figure>\n\n\n\n<figure class=\"wp-block-video\"><video controls src=\"http:\/\/dmslab.hkg03.bdysite.com\/wp-content\/uploads\/2021\/12\/\u9707-2021-12-12-10.14.44.mp4\"><\/video><\/figure>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>Project Members<\/strong><\/h3>\n\n\n\n<ul><li>Zhen Yang <\/li><li>Weitong Chen<\/li><li>Yuting Zhu<\/li><\/ul>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>Publication<\/strong><\/h3>\n\n\n\n<ul><li>Yang Z, Chen W, Huang J. \u201cEnhancing Recommendation on Extremely \n Sparse Data with Blocks-Coupled Non-negative Matrix Factorization.\u201d  \nAccepted by Neurocomputing. <\/li><li>\u9648\u4f1f\u6850. \u9762\u5411\u8d85\u7a00\u758f\u6570\u636e\u7684\u77e9\u9635\u5206\u5757\u8026\u5408\u56e0\u5b50\u5316\u65b9\u6cd5\u7814\u7a76\u4e0e\u5e94\u7528, \u5317\u4eac\u5de5\u4e1a\u5927\u5b66\u7855\u58eb\u5b66\u4f4d\u8bba\u6587\uff0c2017.<\/li><\/ul>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>Code &amp; Toolbox<\/strong><\/h3>\n\n\n\n<ul><li><strong><a href=\"https:\/\/github.com\/fromskyblue\/Extremely-Sparse-Data-Reconstruction\">Github<\/a><\/strong><\/li><\/ul>\n","protected":false},"excerpt":{"rendered":"<p>Extremely Sparse Data with Blocks-Coupled Non-negative  &hellip; <\/p>\n<p class=\"link-more\"><a href=\"http:\/\/www.dmslab.cn\/?page_id=92\" class=\"more-link\">\u7ee7\u7eed\u9605\u8bfb<span class=\"screen-reader-text\">\u201cExtremely Sparse Data\u201d<\/span><\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"parent":62,"menu_order":0,"comment_status":"open","ping_status":"closed","template":"","meta":{"footnotes":""},"_links":{"self":[{"href":"http:\/\/www.dmslab.cn\/index.php?rest_route=\/wp\/v2\/pages\/92"}],"collection":[{"href":"http:\/\/www.dmslab.cn\/index.php?rest_route=\/wp\/v2\/pages"}],"about":[{"href":"http:\/\/www.dmslab.cn\/index.php?rest_route=\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"http:\/\/www.dmslab.cn\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/www.dmslab.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=92"}],"version-history":[{"count":5,"href":"http:\/\/www.dmslab.cn\/index.php?rest_route=\/wp\/v2\/pages\/92\/revisions"}],"predecessor-version":[{"id":1612,"href":"http:\/\/www.dmslab.cn\/index.php?rest_route=\/wp\/v2\/pages\/92\/revisions\/1612"}],"up":[{"embeddable":true,"href":"http:\/\/www.dmslab.cn\/index.php?rest_route=\/wp\/v2\/pages\/62"}],"wp:attachment":[{"href":"http:\/\/www.dmslab.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=92"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}