Collaborative Topic Regression with Social Matrix Factorization for Recommendation Systems

By Sanjay Purushotham, Yan Liu. 2012

195 citations recorded by the source corpus.

Social network websites, such as Facebook, YouTube, Lastfm etc, have become a popular platform for users to connect with each other and share content or opinions. They provide rich information for us to study the influence of user's social circle in their decision process. In this paper, we are interested in examining the effectiveness of social network information to predict the user's ratings of items. We propose a novel hierarchical Bayesian model which jointly incorporates topic modeling and probabilistic matrix factorization of social networks. A major advantage of our model is to automatically infer useful latent topics and social information as well as their importance to collaborative filtering from the training data. Empirical experiments on two large-scale datasets show that our algorithm provides a more effective recommendation system than the state-of-the art approaches. Our results reveal interesting insight that the social circles have more influence on people's decisions about the usefulness of information (e.g., bookmarking preference on Delicious) than personal taste (e.g., music preference on Lastfm). We also examine and discuss solutions on potential information leak in many recommendation systems that utilize social information.

Subfields: Information Systems, Statistical and Nonlinear Physics, Recommender Systems

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Information Systems

In Statistical and Nonlinear Physics

In Recommender Systems

Citations in the corpus

Collaborative Deep Learning for Recommender Systems (2014, methodology): “As in [34, 22, 35], we use recall as the performance measure because the rating information is in the form of implicit”
A Survey of Multi-View Representation Learning (2016, unknown): “[202] propose a hierarchical Bayesian model to connect social network information and item information through shared user latent representation.”
Robust Federated Recommendation System (2020, background): “Recommendation systems (RS), which are machine learning models that try to predict the user (client) preferences, are increasingly being deployed commercially [22; 4; 26; 37; 32; 35].”
Relational Collaborative Topic Regression for Recommender Systems (2015, methodology): “Some other methods [22], [29], [31] try to use social networks among users to improve the performance.”
Ratings meet reviews, a combined approach to recommend (2014, background): “Later work [22] extended it to take the social network among users into consideration.”
Towards Bayesian Deep Learning: A Survey (2016, methodology): “This model is then further extended to incorporate social information [44], [57], [58].”
Convolutional Matrix Factorization for Document Context-Aware Recommendation (2016, background): “2959165 iary information such as demography of users, social networks, and item description documents [14, 15, 17, 19, 23, 24].”
A Survey on Bayesian Deep Learning (2016, unknown): “This model is then further extended to incorporate social information [89, 115, 116].”
A Probabilistic Model for Using Social Networks in Personalized Item Recommendation (2015, methodology): “In the following, we develop the mathematical details behind the model (Section 2), derive an efficient learning algorithm (based on variational inference) for estimating it from data (Section 2, Appendix), and evaluate it on six real-world data sets (Section 3).”
Collaborative Knowledge Base Embedding for Recommender Systems (2016, background): “As discussed in [19], precision is not a suitable performance measure for implicit feedback recommendation.”
Relational Stacked Denoising Autoencoder for Tag Recommendation (2015, methodology): “Recently, models like collaborative topic regression (CTR) (Wang and Blei 2011) and its variants (Purushotham, Liu, and Kuo 2012; Wang, Chen, and Li 2013) have been proposed and adapted for tag recommendation to achieve promising performance.”