1. NCCU Academic Hub
  2. Publications
  3. College of Informatics
  4. Theses
  5. Item

Publications-Theses

Article View/Open

Publication Export

Google ScholarTM

NCCU Library

Citation Infomation

Related Publications in TAIR

題名 基於內容偏好圖卷積網路之完全冷啟動推薦演算法
Improving Complete Cold-Start Recommendation via Content-based Preference Graph Convolution Networks
作者 王韋勝
Wang, Wei-Sheng
貢獻者 蔡銘峰
Tsai, ‪Ming-Feng
王韋勝
Wang, Wei-Sheng
關鍵詞 推薦系統
冷起動推薦
圖學習表示法
Graph Representation
Recommender System
Cold-Start Recommendation
日期 2022
上傳時間 9-Mar-2023 18:37:16 (UTC+8)
摘要 傳統的混合推薦系統旨在結合協同過濾和內容過濾兩種方式進行推 薦，利用使用者喜好資訊和過去互動過的商品內容資訊來解決資料稀 疏性問題和冷啟動問題。但是，在現實世界中，經常因為產品的性質 讓使用者和產品的互動資料相當稀少或是缺少這些資料，從而導致了 完全冷啟動(Complete Cold Start, CCS)問題，如新聞推薦和新活動推 薦，這是傳統的混合模型無法解決的。
在本文中，我們提出了偏好內容卷積(Preference Content Convolu- tion, PCC)方法，這是一種基於圖卷積網絡(Graph Convolution Net- work, GCN)的圖學習表示方法，該方法可在接受缺失資料的前提下 同時抽取使用者對內容的喜好特徵並結合內容資訊，進而針對冷啟動 問題進行推薦。我們在現實世界中的線上售票服務資料集和圖書資料 集上進行的實驗驗證此方法，其性能優於其他傳統基於內容過濾的方 法和沒有卷積網路的混合模型，為基於卷積網路的模型指出了一個方 向。
Conventional hybrid recommender system aims to address the data spar- sity problem and the cold start problem by leveraging collaborative and content- based filtering, simultaneously leveraging the precious user preference infor- mation and staple item content information. However, in many real-world scenarios, such as news and new event recommendations, the nature of items dictates the complete lack of user-item interaction, leading to the complete cold start (CCS) problem, which traditional hybrid models cannot solve.
In this paper, we propose preference-content convolution (PCC), a con- volutional graph network (GCN) based embedding learning method which jointly captures item content information and user preference over item con- tent. The experiments conducted on the real-world online ticket vending service dataset and news recommendation dataset show that the proposed method significantly outperforms traditional content-based filtering methods and hybrid models without convolution, signifying a promising direction for using the convolution-based model in addressing the CCS problem.
參考文獻 [1] A. Bordes, N. Usunier, A. Garcia-Duran, J. Weston, and O. Yakhnenko. Translating embeddings for modeling multi-relational data. Advances in neural information processing systems, 26, 2013.
[2] I. Cantador, A. Bellog ́ın, and D. Vallet. Content-based recommendation in social tagging systems. In Proceedings of the fourth ACM conference on Recommender systems, pages 237–240, 2010.
[3] C.-M. Chen, M.-F. Tsai, Y.-C. Lin, and Y.-H. Yang. Query-based music recommen- dations via preference embedding. In Proceedings of the 10th ACM Conference on Recommender Systems, pages 79–82, 2016.
[4] D. Cohen, M. Aharon, Y. Koren, O. Somekh, and R. Nissim. Expediting exploration by attribute-to-feature mapping for cold-start recommendations. In Proceedings of the Eleventh ACM Conference on Recommender Systems, pages 184–192, 2017.
[5] M. Elahi, F. Ricci, and N. Rubens. Active learning in collaborative filtering rec- ommender systems. In International Conference on Electronic Commerce and Web Technologies, pages 113–124. Springer, 2014.
[6] N. Golbandi, Y. Koren, and R. Lempel. Adaptive bootstrapping of recommender systems using decision trees. In Proceedings of the fourth ACM international con- ference on Web search and data mining, pages 595–604, 2011.
[7] J. Gope and S. K. Jain. A survey on solving cold start problem in recommender systems. In 2017 International Conference on Computing, Communication and Au- tomation (ICCCA), pages 133–138. IEEE, 2017.
[8] W. Hamilton, Z. Ying, and J. Leskovec. Inductive representation learning on large graphs. Advances in neural information processing systems, 30, 2017.
[9] Y. Juan, Y. Zhuang, W.-S. Chin, and C.-J. Lin. Field-aware factorization machines for ctr prediction. In Proceedings of the 10th ACM conference on recommender systems, pages 43–50, 2016.
[10] Y.Koren,R.Bell,andC.Volinsky.Matrixfactorizationtechniquesforrecommender systems. Computer, 42(8):30–37, 2009.
[11] M. Kula. Metadata embeddings for user and item cold-start recommendations. In T. Bogers and M. Koolen, editors, Proceedings of the 2nd Workshop on New Trends on Content-Based Recommender Systems co-located with 9th ACM Conference on Recommender Systems (RecSys 2015), Vienna, Austria, September 16-20, 2015., volume 1448 of CEUR Workshop Proceedings, pages 14–21. CEUR-WS.org, 2015.
[12] B. Lika, K. Kolomvatsos, and S. Hadjiefthymiades. Facing the cold start problem in recommender systems. Expert systems with applications, 41(4):2065–2073, 2014.
[13] P. Melville, R. J. Mooney, R. Nagarajan, et al. Content-boosted collaborative filter- ing for improved recommendations. Aaai/iaai, 23:187–192, 2002.
[14] R. Mihalcea and P. Tarau. Textrank: Bringing order into text. In Proceedings of the 2004 conference on empirical methods in natural language processing, pages 404–411, 2004.
[15] T. Mikolov, K. Chen, G. Corrado, and J. Dean. Efficient estimation of word repre- sentations in vector space. arXiv preprint arXiv:1301.3781, 2013.
[16] B. Perozzi, R. Al-Rfou, and S. Skiena. Deepwalk: Online learning of social repre- sentations. In Proceedings of the 20th ACM SIGKDD international conference on Knowledge discovery and data mining, pages 701–710, 2014.
[17] S. Rendle, C. Freudenthaler, Z. Gantner, and L. Schmidt-Thieme. Bpr: Bayesian personalized ranking from implicit feedback. arXiv preprint arXiv:1205.2618, 2012.
[18] B. Sarwar, G. Karypis, J. Konstan, and J. Riedl. Item-based collaborative filtering recommendation algorithms. In Proceedings of the 10th international conference on World Wide Web, pages 285–295, 2001.
[19] M. Saveski and A. Mantrach. Item cold-start recommendations: learning local col- lective embeddings. In Proceedings of the 8th ACM Conference on Recommender systems, pages 89–96, 2014.
[20] B. Shapira, L. Rokach, and S. Freilikhman. Facebook single and cross domain data for recommendation systems. User Modeling and User-Adapted Interaction, 23(2):211–247, 2013.
[21] X. Su and T. M. Khoshgoftaar. A survey of collaborative filtering techniques. Ad- vances in artificial intelligence, 2009, 2009.
[22] M. Sun, F. Li, J. Lee, K. Zhou, G. Lebanon, and H. Zha. Learning multiple-question decision trees for cold-start recommendation. In Proceedings of the sixth ACM in- ternational conference on Web search and data mining, pages 445–454, 2013.
[23] J. Tang, M. Qu, M. Wang, M. Zhang, J. Yan, and Q. Mei. Line: Large-scale infor- mation network embedding. In Proceedings of the 24th international conference on world wide web, pages 1067–1077, 2015.
[24] H. Wang, M. Zhao, X. Xie, W. Li, and M. Guo. Knowledge graph convolutional net- works for recommender systems. In The world wide web conference, pages 3307– 3313, 2019.
[25] J. Wei, J. He, K. Chen, Y. Zhou, and Z. Tang. Collaborative filtering and deep learning based recommendation system for cold start items. Expert Systems with Applications, 69:29–39, 2017.
[26] K. Zhou, S.-H. Yang, and H. Zha. Functional matrix factorizations for cold-start recommendation. In Proceedings of the 34th international ACM SIGIR conference on Research and development in Information Retrieval, pages 315–324, 2011.
描述 碩士
國立政治大學
資訊科學系
109753110
資料來源 http://thesis.lib.nccu.edu.tw/record/#G0109753110
資料類型 thesis
dc.contributor.advisor 蔡銘峰zh_TW
dc.contributor.advisor Tsai, ‪Ming-Fengen_US
dc.contributor.author (Authors) 王韋勝zh_TW
dc.contributor.author (Authors) Wang, Wei-Shengen_US
dc.creator (作者) 王韋勝zh_TW
dc.creator (作者) Wang, Wei-Shengen_US
dc.date (日期) 2022en_US
dc.date.accessioned 9-Mar-2023 18:37:16 (UTC+8)-
dc.date.available 9-Mar-2023 18:37:16 (UTC+8)-
dc.date.issued (上傳時間) 9-Mar-2023 18:37:16 (UTC+8)-
dc.identifier (Other Identifiers) G0109753110en_US
dc.identifier.uri (URI) http://nccur.lib.nccu.edu.tw/handle/140.119/143835-
dc.description (描述) 碩士zh_TW
dc.description (描述) 國立政治大學zh_TW
dc.description (描述) 資訊科學系zh_TW
dc.description (描述) 109753110zh_TW
dc.description.abstract (摘要) 傳統的混合推薦系統旨在結合協同過濾和內容過濾兩種方式進行推 薦，利用使用者喜好資訊和過去互動過的商品內容資訊來解決資料稀 疏性問題和冷啟動問題。但是，在現實世界中，經常因為產品的性質 讓使用者和產品的互動資料相當稀少或是缺少這些資料，從而導致了 完全冷啟動(Complete Cold Start, CCS)問題，如新聞推薦和新活動推 薦，這是傳統的混合模型無法解決的。
在本文中，我們提出了偏好內容卷積(Preference Content Convolu- tion, PCC)方法，這是一種基於圖卷積網絡(Graph Convolution Net- work, GCN)的圖學習表示方法，該方法可在接受缺失資料的前提下 同時抽取使用者對內容的喜好特徵並結合內容資訊，進而針對冷啟動 問題進行推薦。我們在現實世界中的線上售票服務資料集和圖書資料 集上進行的實驗驗證此方法，其性能優於其他傳統基於內容過濾的方 法和沒有卷積網路的混合模型，為基於卷積網路的模型指出了一個方 向。		zh_TW
dc.description.abstract (摘要) Conventional hybrid recommender system aims to address the data spar- sity problem and the cold start problem by leveraging collaborative and content- based filtering, simultaneously leveraging the precious user preference infor- mation and staple item content information. However, in many real-world scenarios, such as news and new event recommendations, the nature of items dictates the complete lack of user-item interaction, leading to the complete cold start (CCS) problem, which traditional hybrid models cannot solve.
In this paper, we propose preference-content convolution (PCC), a con- volutional graph network (GCN) based embedding learning method which jointly captures item content information and user preference over item con- tent. The experiments conducted on the real-world online ticket vending service dataset and news recommendation dataset show that the proposed method significantly outperforms traditional content-based filtering methods and hybrid models without convolution, signifying a promising direction for using the convolution-based model in addressing the CCS problem.		en_US
dc.description.tableofcontents Ch1 緒論 1
1.1 前言 1
1.2 研究目的 2

Ch 2 相關文獻探討 5
2.1 推薦系統(RecommenderSystem) 5
2.1.1 內容過濾(ContentFiltering) 5
2.1.2 協同過濾(CollaborativeFiltering) 6
2.2 冷啟動問題(ColdStartProblem) 9
2.3 圖學習表示法(GraphRepresentation) 10

Ch3 研究方法 12
3.1 問題定義 12
3.2 偏好內容圖卷積框架 13
3.2.1 建圖策略 13
3.2.2 偏好空間建構(Preference Space Construction) 15
3.2.3 內容空間建構(Content Space Construction) 16
3.2.4 混合空間建構(Hybrid Space Construction) 18

Ch4 實驗結果與討論 21
4.1 資料集 21
4.1.1 資料前處理 22
4.2 比較基準模型 22
4.3 實驗設定與評估標準 23
4.3.1 實驗設定 23
4.3.2 評估標準 23
4.4 實驗結果 25
4.4.1 完全冷啟動推薦 25
4.4.2 商品對商品推薦(Item to Item Recommendation) 26
4.5 實例分析 27

Ch5 結論 30
參考文獻 31		zh_TW
dc.format.extent 6624716 bytes-
dc.format.mimetype application/pdf-
dc.source.uri (資料來源) http://thesis.lib.nccu.edu.tw/record/#G0109753110en_US
dc.subject (關鍵詞) 推薦系統zh_TW
dc.subject (關鍵詞) 冷起動推薦zh_TW
dc.subject (關鍵詞) 圖學習表示法zh_TW
dc.subject (關鍵詞) Graph Representationen_US
dc.subject (關鍵詞) Recommender Systemen_US
dc.subject (關鍵詞) Cold-Start Recommendationen_US
dc.title (題名) 基於內容偏好圖卷積網路之完全冷啟動推薦演算法zh_TW
dc.title (題名) Improving Complete Cold-Start Recommendation via Content-based Preference Graph Convolution Networksen_US
dc.type (資料類型) thesisen_US
dc.relation.reference (參考文獻) [1] A. Bordes, N. Usunier, A. Garcia-Duran, J. Weston, and O. Yakhnenko. Translating embeddings for modeling multi-relational data. Advances in neural information processing systems, 26, 2013.
[2] I. Cantador, A. Bellog ́ın, and D. Vallet. Content-based recommendation in social tagging systems. In Proceedings of the fourth ACM conference on Recommender systems, pages 237–240, 2010.
[3] C.-M. Chen, M.-F. Tsai, Y.-C. Lin, and Y.-H. Yang. Query-based music recommen- dations via preference embedding. In Proceedings of the 10th ACM Conference on Recommender Systems, pages 79–82, 2016.
[4] D. Cohen, M. Aharon, Y. Koren, O. Somekh, and R. Nissim. Expediting exploration by attribute-to-feature mapping for cold-start recommendations. In Proceedings of the Eleventh ACM Conference on Recommender Systems, pages 184–192, 2017.
[5] M. Elahi, F. Ricci, and N. Rubens. Active learning in collaborative filtering rec- ommender systems. In International Conference on Electronic Commerce and Web Technologies, pages 113–124. Springer, 2014.
[6] N. Golbandi, Y. Koren, and R. Lempel. Adaptive bootstrapping of recommender systems using decision trees. In Proceedings of the fourth ACM international con- ference on Web search and data mining, pages 595–604, 2011.
[7] J. Gope and S. K. Jain. A survey on solving cold start problem in recommender systems. In 2017 International Conference on Computing, Communication and Au- tomation (ICCCA), pages 133–138. IEEE, 2017.
[8] W. Hamilton, Z. Ying, and J. Leskovec. Inductive representation learning on large graphs. Advances in neural information processing systems, 30, 2017.
[9] Y. Juan, Y. Zhuang, W.-S. Chin, and C.-J. Lin. Field-aware factorization machines for ctr prediction. In Proceedings of the 10th ACM conference on recommender systems, pages 43–50, 2016.
[10] Y.Koren,R.Bell,andC.Volinsky.Matrixfactorizationtechniquesforrecommender systems. Computer, 42(8):30–37, 2009.
[11] M. Kula. Metadata embeddings for user and item cold-start recommendations. In T. Bogers and M. Koolen, editors, Proceedings of the 2nd Workshop on New Trends on Content-Based Recommender Systems co-located with 9th ACM Conference on Recommender Systems (RecSys 2015), Vienna, Austria, September 16-20, 2015., volume 1448 of CEUR Workshop Proceedings, pages 14–21. CEUR-WS.org, 2015.
[12] B. Lika, K. Kolomvatsos, and S. Hadjiefthymiades. Facing the cold start problem in recommender systems. Expert systems with applications, 41(4):2065–2073, 2014.
[13] P. Melville, R. J. Mooney, R. Nagarajan, et al. Content-boosted collaborative filter- ing for improved recommendations. Aaai/iaai, 23:187–192, 2002.
[14] R. Mihalcea and P. Tarau. Textrank: Bringing order into text. In Proceedings of the 2004 conference on empirical methods in natural language processing, pages 404–411, 2004.
[15] T. Mikolov, K. Chen, G. Corrado, and J. Dean. Efficient estimation of word repre- sentations in vector space. arXiv preprint arXiv:1301.3781, 2013.
[16] B. Perozzi, R. Al-Rfou, and S. Skiena. Deepwalk: Online learning of social repre- sentations. In Proceedings of the 20th ACM SIGKDD international conference on Knowledge discovery and data mining, pages 701–710, 2014.
[17] S. Rendle, C. Freudenthaler, Z. Gantner, and L. Schmidt-Thieme. Bpr: Bayesian personalized ranking from implicit feedback. arXiv preprint arXiv:1205.2618, 2012.
[18] B. Sarwar, G. Karypis, J. Konstan, and J. Riedl. Item-based collaborative filtering recommendation algorithms. In Proceedings of the 10th international conference on World Wide Web, pages 285–295, 2001.
[19] M. Saveski and A. Mantrach. Item cold-start recommendations: learning local col- lective embeddings. In Proceedings of the 8th ACM Conference on Recommender systems, pages 89–96, 2014.
[20] B. Shapira, L. Rokach, and S. Freilikhman. Facebook single and cross domain data for recommendation systems. User Modeling and User-Adapted Interaction, 23(2):211–247, 2013.
[21] X. Su and T. M. Khoshgoftaar. A survey of collaborative filtering techniques. Ad- vances in artificial intelligence, 2009, 2009.
[22] M. Sun, F. Li, J. Lee, K. Zhou, G. Lebanon, and H. Zha. Learning multiple-question decision trees for cold-start recommendation. In Proceedings of the sixth ACM in- ternational conference on Web search and data mining, pages 445–454, 2013.
[23] J. Tang, M. Qu, M. Wang, M. Zhang, J. Yan, and Q. Mei. Line: Large-scale infor- mation network embedding. In Proceedings of the 24th international conference on world wide web, pages 1067–1077, 2015.
[24] H. Wang, M. Zhao, X. Xie, W. Li, and M. Guo. Knowledge graph convolutional net- works for recommender systems. In The world wide web conference, pages 3307– 3313, 2019.
[25] J. Wei, J. He, K. Chen, Y. Zhou, and Z. Tang. Collaborative filtering and deep learning based recommendation system for cold start items. Expert Systems with Applications, 69:29–39, 2017.
[26] K. Zhou, S.-H. Yang, and H. Zha. Functional matrix factorizations for cold-start recommendation. In Proceedings of the 34th international ACM SIGIR conference on Research and development in Information Retrieval, pages 315–324, 2011.		zh_TW