AI帮你理解科学

AI 生成解读视频

AI抽取解析论文重点内容自动生成视频


pub
生成解读视频

AI 溯源

AI解析本论文相关学术脉络


Master Reading Tree
生成 溯源树

AI 精读

AI抽取本论文的概要总结


微博一下
High performance of matrix factorization method increases the accuracy of predicted rating in the user-items rating matrix

ISoTrustSeq: a social recommender system based on implicit interest, trust and sequential behaviors of users using matrix factorization

Journal of Intelligent Information Systems, no. 2.0 (2019): 1-30

被引用2|浏览2
EI
下载 PDF 全文
引用
微博一下

摘要

Recommender systems try to propose a list of main interests of an on line social network user based on his predicted rating values. In the recent years, several methods are proposed such as Interest Social Recommendation method (ISoRec), and Social Recommendation method based on trust Sequence Matrix Factorization which employs matrix fac...更多

代码

数据

0
简介
  • Recommender systems suggest a list of items that a user may be interested in.
  • These systems help users to choosed their suitable items in a very short time.
  • There are two different categories of the Recommender systems: Context base and Collaborative filtering (CF) (Bobadilla et al 2013).
  • Among the several types of recommender systems, collaborative filtering is one of the most popular methods which is simple to implement and can discover complicated patterns (Zhang et al 2013)
重点内容
  • Recommender systems suggest a list of items that a user may be interested in
  • Collaborative filtering-based recommender systems are divided into two categories: Matrix Factorization based social recommender methods (SocialMF) and Neighborhood based social recommender methods (Su and Khoshgoftaar 2009)
  • High performance of matrix factorization method increases the accuracy of predicted rating in the user-items rating matrix
  • We proposed the conditional probability distribution in order to reflect the impact of the relationship between implicit user’ interests and social trust of the users on their judgment in relation to the items
  • Users’ rating values include integers from one to five which shows the user interest to the desired item According to type of data, the value but will affect the decision making by other customers
  • We examine the performance of the proposed algorithm compared with two Interest Social Recommendation method (ISoRec) and TrustSeqMF methods by changes in the number of user and items features
结果
  • Used dataset in this study is Epinions. Epinions.com is a very popular review site that was established in 1999.
  • Epinions dataset rating matrix contains four columns
  • The content of this column includes the user ID number, item code, the useritem rating and the temporal of rating.
  • Another Matrix of this data set has kept information on the users’ trust in itself as trust network.
  • The content of each of the table cells represents the root mean square error (RMSE)
结论
  • Conclusion and future work

    Implicit Social Trust Sequence (ISoTrustSeq) method is presented to predict missing ratings of the users in the user-item rating matrix.
总结
  • Introduction:

    Recommender systems suggest a list of items that a user may be interested in.
  • These systems help users to choosed their suitable items in a very short time.
  • There are two different categories of the Recommender systems: Context base and Collaborative filtering (CF) (Bobadilla et al 2013).
  • Among the several types of recommender systems, collaborative filtering is one of the most popular methods which is simple to implement and can discover complicated patterns (Zhang et al 2013)
  • Objectives:

    The main goal of this paper is to solve user-item rating based on the trust, sequential interest and the implicit interest of users, simultaneously.
  • The aim of this study is to provide a recommender system which can achieve the users, directors and producers’ satisfaction by providing some recommendations comply with or close to their views
  • Results:

    Used dataset in this study is Epinions. Epinions.com is a very popular review site that was established in 1999.
  • Epinions dataset rating matrix contains four columns
  • The content of this column includes the user ID number, item code, the useritem rating and the temporal of rating.
  • Another Matrix of this data set has kept information on the users’ trust in itself as trust network.
  • The content of each of the table cells represents the root mean square error (RMSE)
  • Conclusion:

    Conclusion and future work

    Implicit Social Trust Sequence (ISoTrustSeq) method is presented to predict missing ratings of the users in the user-item rating matrix.
表格
  • Table1: The comparison of proposed method (ISoTrustSeq) with ISoRec, TrustSeqMF and SocialMF K ∈ {2, 5}
  • Table2: The comparison of proposed method (ISoTrustSeq) with ISoRec, TrustSeqMF and SocialMF K ∈ {10, 15}
  • Table3: The comparison of proposed method (ISoTrustSeq) with ISoRec, TrustSeqMF and SocialMF K ∈ {20, 27}
Download tables as Excel
基金
  • In the collaborative filtering, high performance of matrix factorization method increases the accuracy of predicted rating in the user-items rating matrix
  • The results of the comparisons made with two ISoRec and TrustSeqMF methods show that the proposed method increases the prediction accuracy more than other methods
研究对象与分析
users: 4
Figure 11 shows a scenario of the real-world recommendations including three central elements, users’ social trust network, item consumption network and user-item rating matrix. In this scenario, there are four relationships between 4 users (u1 to u4) and a relationships between two items (v1 to v2). Prediction of the missing values ru,i ,(the rating of the user u to the item i) through the user-item rating matrix R, trust matrix T and items consumption matrix S, are the task of recommenders

relationships between five users: 6
Each node represents a user and each edge represents the trusted relationship between them. As seen in the figure, the graph of trust between users consists of six relationships between five users (u1 to u5 ∈ U ). T = {tij } shows a N × N matrix of the user’s trust graph and are called social trust matrix

引用论文
  • Bobadilla, J., Ortega, F., Hernando, A., Gutierrez, A. (2013). Recommender systems survey. KnowledgeBased Systems, 46, 109–132.
    Google ScholarLocate open access versionFindings
  • Bottou, L. (2010). Large-scale machine learning with stochastic gradient descent. In Proceedings of the International Conference on Computer Statistics.
    Google ScholarLocate open access versionFindings
  • Ford, J., Makedon, F., Pearlman, J. (2005). Using singular value decomposition approximation for collaborative filtering. In 7th IEEE International Conference on E-Commerce Technology (pp. 257– 264).
    Google ScholarLocate open access versionFindings
  • Han, J., Kamber, M., Pei, J. (2006). Data mining – consepts and techniques, edition S. San Francisco: Morgan Kaufmann of Elsevier.
    Google ScholarFindings
  • Hofmann, T. (2003). Collaborative filtering via gaussian probabilistic latent semantic analysis. In Proceedings of the 26th Annual International ACM SIGIR Conference on Research and Development Informaion Retrieval (pp. 259–266).
    Google ScholarLocate open access versionFindings
  • Jamali, M. (2010). A Matrix Factorization Technique with Trust Propagation for Recommendation in Social Networks Categories and Subject Descriptors. In The fourth ACM conference on Recommender systems (pp. 135–142). no. 978–1–60558–906–0.
    Google ScholarLocate open access versionFindings
  • Kadie, C., Breese, J.S., Heckerman, D. (1998). Empirical analysis of predictive algorithms for collaborative filtering. In UAI-Fourteenth conference on Uncertainty in artificial intelligence (pp. 43–52).
    Google ScholarLocate open access versionFindings
  • Kannan, R., Ishteva, M., Park, H. (2013). Bounded matrix factorization for recommender system. Knowledge Information Systems, 39(3), 491–511.
    Google ScholarLocate open access versionFindings
  • Koren, Y., Bell, R., Volinsky, C. (2009). Matrix factorization techniques for recommender systems. Computer (Long Beach California), 42(8), 30–37.
    Google ScholarLocate open access versionFindings
  • Li, B., Tian, B.-B., Liu, J. (2015). A Novel Probabilistic Matrix Factorization Recommendation Algorithm Using Social Reliable Similarity Propagation. Intelligence Computing Theory Methodology, 9226, 80– 91.
    Google ScholarLocate open access versionFindings
  • Ma, H., Yang, H., Lyu, M.R., King, I. (2008). SoRec: Social recommendation using probabilistic matrix factorization. In Proceeding of the 17th ACM Conference on Information Knowledge Management vol. 08 (pp. 0–9).
    Google ScholarLocate open access versionFindings
  • Ma, H., King, I., Lyu, M.R. (2011). Learning to Recommend with Social Trust Ensemble. In SIGIR ’09 Proceedings of the 32nd International ACM SIGIR Conference (pp. 203–210).
    Google ScholarLocate open access versionFindings
  • Ma, H., Zhou, D., Liu, C., Lyu, M.R. (2011). Recommender systems with social regularization. In Proceedings of the fourth ACM International Conference on Web Search Data Mining, no. 978–1–4503–0493–1 (pp. 287–296).
    Google ScholarLocate open access versionFindings
  • Matuszyk, P., Vinagre, J., Spiliopoulou, M., Jorge, A.M., Gama, J. (2015). Forgetting methods for incremental matrix factorization in recommender systems. In Proceedings of the 30th Annual ACM Symposium on Applied Computing - SAC ’15 (pp. 947–953).
    Google ScholarLocate open access versionFindings
  • Oh, J., & Han, W. (2015). Fast and robust parallel SGD matrix factorization. In KDD ’15 Proceedings of the 21th ACM SIGKDD, The International Conference on Knowledge Discovery and Data Mining (pp. 865– 874).
    Google ScholarLocate open access versionFindings
  • Qian, X., Feng, H., Zhao, G., Mei, T., Member, S. (2014). Personalized recommendation combining user interest and social circle. IEEE Transactions on Knowledge and Data Engineering, 26(7), 1763–1777.
    Google ScholarLocate open access versionFindings
  • Rana, C., & Jain, S.K. (2014). An evolutionary clustering algorithm based on temporal features for dynamic recommender systems. Swarm and Evolutionary Computation, 14, 21–30.
    Google ScholarLocate open access versionFindings
  • Recht, R.C.B. (2013). Parallel stochastic gradient algorithms for large-scale matrix completion. Mathematical Programming Computation, 5, 201–226.
    Google ScholarLocate open access versionFindings
  • Salakhutdinov, R., & Mnih, A. (2008). Probabilistic matrix factorization. Neural Information Processing System, 1–8.
    Google ScholarLocate open access versionFindings
  • Su, X., & Khoshgoftaar, T.M. (2009). A survey of collaborative filtering techniques. Advances in Artificial Intelligence, 2009(3), 1–19.
    Google ScholarLocate open access versionFindings
  • Way, O.M. (2013). An Experimental Study on Implicit Social Recommendation. In 36th international ACM SIGIR conference on Research and development in information retrieval (pp. 73–82).
    Google ScholarLocate open access versionFindings
  • Wu, L., Chen, E., Liu, Q., Xu, L., Bao, T., Zhang, L. (2012). Leveraging tagging for neighborhood-aware probabilistic matrix factorization. In 21st ACM international conference on Information and knowledge management (pp. 1854–1858).
    Google ScholarLocate open access versionFindings
  • Yang, X., Member, S., Guo, Y., Liu, Y. (2013). Bayesian-inference-based recommendation in online social networks. IEEE Transactions on Parallel and Distributed Systems, 24(4), 642–651.
    Google ScholarLocate open access versionFindings
  • Yang, X., Guo, Y., Liu, Y., Steck, H. (2014). A survey of collaborative filtering based social recommender systems. Computer Communications, 41, 1–10.
    Google ScholarLocate open access versionFindings
  • Yu, H.-F., Hsieh, C.-J., Si, S., Dhillon, I.S. (2013). Parallel matrix factorization for recommender systems. Knowledge Information Systems, 41(3), 793–819.
    Google ScholarLocate open access versionFindings
  • Yu-sheng, L.I., Mei-na, S., Jun-de, S. (2014). Social recommendation algorithm fusing user interest social network. The Journal of China Universities of Posts and Telecommunications, 21, 26–33.
    Google ScholarLocate open access versionFindings
  • Zhang, Z., & Liu, H. (2014). Application and research of improved probability matrix factorization techniques in collaborative filtering. International Journal of Control, Automation and Systems, 7(8), 79– 92.
    Google ScholarLocate open access versionFindings
  • Zhang, Z., & Liu, H. (2015). Social recommendation model combining trust propagation and sequential behaviors. Applied Intelligence, 43(3), 695–706.
    Google ScholarLocate open access versionFindings
  • Zhang, C., Zhang, Z., Yu, L., Liu, C., Liu, H. (2013). Information filtering via collaborative user clustering modeling. Physica A, 396, 195–203.
    Google ScholarLocate open access versionFindings
  • Zhang, Y., Chen, M., Huang, D., Wu, D., Li, Y. (2016). iDoctor: Personalized and professionalized medical recommendations based on hybrid matrix factorization, Future Generation Computer Systems, p. In Press.
    Google ScholarFindings
作者
Vahideh Nobahari
Vahideh Nobahari
Seyyed Javad Seyyed Mahdavi
Seyyed Javad Seyyed Mahdavi
您的评分 :
0

 

标签
评论
数据免责声明
页面数据均来自互联网公开来源、合作出版商和通过AI技术自动分析结果,我们不对页面数据的有效性、准确性、正确性、可靠性、完整性和及时性做出任何承诺和保证。若有疑问,可以通过电子邮件方式联系我们:report@aminer.cn
小科