Personalized recommender systems based on social relationships and historical behaviors

https://doi.org/10.1016/j.amc.2022.127549Get rights and content

Abstract

Recommender systems have a wide range of applications in the age suffering information overload. A promising way to design better recommender systems in the presence of ubiquitous social media is to utilize social relationships in recommendation algorithms, named social recommendation. One critical challenge in social recommendation is how to mine valuable information intrinsic to social relationships and integrate such information into the algorithm design. In this paper, we argue that both social relationships and historical behaviors are affected by the same implicit factors. For example, due to the existence of implicit factors such as peer influence or common interests in social networks, users with similar implicit factors will have a high probability to become friends and collect similar objects. Accordingly, we propose a recommendation algorithm that jointly utilizes social relationships and historical behaviors, based on the extended linear optimization technique. We test the performance of our algorithm for four groups of users on real networks, including all users, active users, inactive users and cold-start users. Results show that, in all the above four scenarios, the proposed algorithm performs overall best subject to accuracy and diversity metrics compared with the benchmarks. In particular, the algorithm remarkably improves the recommendation performance for cold-start users. Further analysis shows that the contribution of social relationships depends on the coupling strength between social relationships and historical behaviors.

Introduction

With the rapid development of Web 2.0, people enjoy the convenience it brings while also suffering from the information overload dilemma. We have to spend more and more time finding out information that we are interested in. It not only degrades our surfing experience but also reduces social productivity [1]. Driven by advances in computer sciences [2], [3] and social physics [4], [5], personalized recommender systems come into being with the mission of solving this dilemma. As a specialized tool, personalized recommender systems [6], [7], [8], [9], [10] employ users’ historical behaviors, personal profiles, tag information and other relevant information to recommend objects to users. One of the keys to achieve good recommendation performance is to find out the interaction mechanism between users and objects as accurately as possible. If we can reveal mechanisms intrinsic to interactions between users and objects, we can design a well-performing recommendation algorithm. Some people choose one object because the object is recommended by their friend. However, Some people may unknowingly choose the same objects as strangers because they have the same taste. The latter is the underlying idea of the well-known collaborative filtering algorithms [11], [12]. That is, similar users tend to select similar objects. However, classical collaborative filtering algorithms ignore the influence of social relationships on users’ historical behaviors, and it is practically impossible to obtain the social influence only by considering users’ historical behaviors. This problem exists in classical physical dynamics algorithms as well [13], [14].

In recent years, the booming of social media makes it possible to acquire social influence. The commercial value of social media has become a new source of economic growth [15], [16], [17]. Many studies [18], [19], [20] show that social media can improve the recommendation performance. The underlying reasons are twofold: on the one hand, homophily theory [21] indicates that users with social relationships tend to share similar consumer tastes or behaviors. For example, obese persons in a social network tend to gather together [22], and smokers or nonsmokers in a social network also tend to gather together [23]. On the other hand, users’ reviews and consumer opinions for objects on social media will influence their neighbors’ purchase decisions and attitudes toward the objects. The phenomena has been revealed by extensive empirical studies: a user’s purchase probability to a DVD increases with the number of incoming recommendations from his friends [24]; The communication between buyers in Taobao (a consumer marketplace in China) will largely drive the purchasing activity of users [25]; Users with social relationships in an Asian mobile network tend to make similar purchase decisions in purchasing caller-back tones [26]; Connected users in Cyworld (an online social networking site in Korea) will be positively affected by their friends’ purchase behaviors [27]; The purchase willingness of users in QQ (a social network in China) is more likely influenced by the number of prior adoptions in their neighborhoods than the well-connected neighbors [28]. In summary, social relationships are closely related to users’ historical behaviors.

However, not all social relationships are directly relevant to users’ historical behaviors. Here we take FriendFeed and Epinions as examples (see Section 4.1 for detailed descriptions of datasets). We show the corresponding data structure in Fig. 1(a). Namely, interactions take place among users in the user layer, and also take place between the user layer and the object layer. Overall speaking, users with direct social relationships tend to interact with more objects than users without direct social relationships (see Fig. 1(b) and (c)), and strongly similar users in social networks tend to interact with more objects than weakly similar users (see Fig. 1(d)). Taking a close look, one can find that behavioral conversion rates brought by social relationships vary widely. The behavioral conversion rate here is defined as the contribution from one user to its immediate neighbors. For example, in Fig. 1(a), objects α2, α3 and α4 are collected by user i4, and objects α3, α4, α5 and α6 are collected by the neighbor i6 of user i4. Thus, the behavioral conversion rate from user i6 to user i4 is hi6i4=|{α2,α3,α4}{α3,α4,α5,α6}||{α2,α3,α4}|=23, and the behavioral conversion rate from user i4 to user i6 is hi4i6=|{α2,α3,α4}{α3,α4,α5,α6}||{α3,α4,α5,α6}|=12. We show the distribution of behavioral conversion rates of users on FriendFeed and Epinions in Fig. 1(e) and (f). On FriendFeed and Epinions, social relationships with behavioral conversion rates larger than 0.2 separately account for 7.84% and 2.12%, and social relationships with behavioral conversion rates of 0 separately account for 53.67% and 45.49%. In other words, many of the explicit social relationships are not directly related to users’ historical behaviors. This inspires us to pay much more attention to increase the strength of social relationships related to users’ historical behaviors and decrease the strength of social relationships unrelated to or weakly related to users’ historical behaviors.

To this end, we assume that valuable social relationships and historical behaviors are affected by the same implicit factors, thus contributions of the two types of relations can be jointly constrained by each other through implicit factors. We employ the linear optimization technique [29] to obtain contribution weights of implicit factors through joint constraints from social relationships and historical behaviors (i.e., user-object interactions), and then use the weights to make recommendations. We name this algorithm as LOS (short for Linear Optimization for Social recommendation, source codes and datasets can be downloaded from https://github.com/Sarahlibra/LOS). To test the performance of LOS, we conduct experiments for four groups of users including all users, active users (users with enough historical behaviors), inactive users (users with a small number of historical behaviors) and cold-start users (users without historical behaviors). Results show that LOS outperforms 6 benchmarks subject to accuracy metrics, and LOS remarkably enhances the recommendation accuracy for inactive users and cold-start users. To our surprise, LOS is competitive with algorithms dedicated to the diversity or accuracy-diversity dilemma.

Section snippets

Related works

The emergence of social media promotes social relationships to be widely used in recommender systems. Many effective algorithms that consider social relationships are proposed including matrix factorization algorithms, probabilistic-based algorithms and others [30], [31], [32], [33], [34], [35], [36], [37], [38], [39]. These earlier social recommendation algorithms need extra information (such as users’ rating records, users’ profiles, domain knowledge) to predict interactions between users and

Methods

In this section, we first describe the formalized process of the proposed algorithm LOS, and then show the corresponding optimized process.

Datasets

FriendFeed and Epinions [41] are used to test the performance of LOS. FriendFeed (http://www.friendfeed.com/) is a real-time feed aggregation website that updates the information from media like Twitter, YouTube and Delicious, which contains follower-followee relationships and rating relationships between users and objects. Epinions (http://www.epinions.com/) is a consumer review website, which contains trust relationships and rating relationships between users and objects. Social networks are

Analysis

There is a general phenomenon: compared with algorithms that only consider historical behaviors, algorithms that also consider social relationships can significantly improve the recommendation performance for some datasets, while for some other datasets the improvement is non-significant. This promotes us to answer how much do social relationships in LOS contribute.

We first compare the recommendation performance of LOS with its’ degenerated algorithm subject to AUPR for all users. We can obtain

Conclusion and discussion

Social network is a great fortune for recommender systems. How to make better use of it remains a challenge. This paper presents an alternative solution to this challenge: through joint constraints from social networks and user-object interaction networks, we can obtain social relationships that are valuable for recommender systems. While social relationships will does not work if there is no relevance between social networks and user-object interaction networks. In this case, the performance

Acknowledgments

This work was partially supported by the Key Scientific Research Fund of Xihua University under Grant No. Z222022, the Ministry of Education of Humanities and Social Science Project under Grant No. 21JZD055,and the National Natural Science Foundation of China (Grant Nos. 11975071 and 61802316).

References (75)

  • L.-J. Chen et al.

    A trust-based recommendation method using network diffusion processes

    Physica A

    (2018)
  • X. Deng et al.

    A general and effective diffusion-based recommendation scheme on coupled social networks

    Inf. Sci.

    (2017)
  • E. Rich

    User modeling via stereotypes

    Cogn. Sci.

    (1979)
  • Z.-K. Zhang et al.

    Personalized recommendation via integrated diffusion on user–item–tag tripartite graphs

    Physica A

    (2010)
  • S. Xu et al.

    Recommending investors for new startups by integrating network diffusion and investors’ domain preference

    Inf. Sci.

    (2020)
  • L.-J. Chen et al.

    A vertex similarity index for better personalized recommendation

    Physica A

    (2017)
  • W. Wang et al.

    Generalized k-core percolation on higher-order dependent networks

    Appl. Math. Comput.

    (2022)
  • W. Li et al.

    Competing spreading dynamics in simplicial complex

    Appl. Math. Comput.

    (2022)
  • P. Hemp

    Death by information overload

    Harv. Bus. Rev.

    (2009)
  • M.I. Jordan et al.

    Machine learning: trends, perspectives, and prospects

    Science

    (2015)
  • Y. LeCun et al.

    Deep learning

    Nature

    (2015)
  • M. Perc

    The social physics collective

    Sci. Rep.

    (2019)
  • B. Hui et al.

    Personalized recommendation system based on knowledge embedding and historical behavior

    Appl. Intell.

    (2022)
  • H. Gao et al.

    Mining consuming behaviors with temporal evolution for personalized recommendation in mobile marketing apps

    Mob. Netw. Appl.

    (2020)
  • J.B. Schafer et al.

    Collaborative Filtering Recommender Systems

    (2007)
  • Y. Koren et al.

    Advances in collaborative filtering

    (2015)
  • T. Zhou et al.

    Bipartite network projection and personal recommendation

    Phys. Rev. E

    (2007)
  • Y.-C. Zhang et al.

    Heat conduction process on community networks as a recommendation model

    Phys. Rev. Lett.

    (2007)
  • H. Chen et al.

    Exploring the commercial value of social networks: enhancing consumers’ brand experience through facebook pages

    J. Enterp. Inf. Manag.

    (2014)
  • K. Hutter et al.

    The impact of user interactions in social media on brand awareness and purchase intention: the case of mini on facebook

    J. Prod. Brand Manag.

    (2013)
  • O. Arazy et al.

    Improving social recommender systems

    IT Prof.

    (2009)
  • P. Symeonidis et al.

    Product recommendation and rating prediction based on multi-modal social networks

    Proceedings of the 15th ACM Conference on Recommender Systems

    (2011)
  • H. Zheng et al.

    Personalized fashion recommendation from personal social media data: an item-to-set metric learning approach

    Proceedings of 2021 International Conference on Big Data

    (2021)
  • M. McPherson et al.

    Birds of a feather: homophily in social networks

    Annu. Rev. Sociol.

    (2001)
  • N.A. Christakis et al.

    The spread of obesity in a large social network over 32 years

    N. Engl. J. Med.

    (2007)
  • N.A. Christakis et al.

    The collective dynamics of smoking in a large social network

    N. Engl. J. Med.

    (2008)
  • J. Leskovec et al.

    The dynamics of viral marketing

    ACM Trans. Web

    (2007)
  • Cited by (8)

    View all citing articles on Scopus
    View full text