In this paper, we propose a novel spatio-temporal model for collaborative filtering applications. Our model is based on low-rank matrix factorization that uses a spatio-temporal filtering approach to estimate user and item factors. The spatial component regularizes the factors by exploiting correlation across users and/or items, modeled as a function of some implicit feedback (e.g., who rated what) and/or some side information (e.g., user demographics, browsing history). In particular, we incorporate correlation in factors through a Markov random field prior in a probabilistic framework, whereby the neighborhood weights are functions of user and item covariates. The temporal component ensures that the user/item factors adapt to process changes that occur through time and is implemented in a state space framework with fast estimation through Kalman filtering. Our spatio-temporal filtering (ST-KF hereafter) approach provides a single joint model to simultaneously incorporate both spatial and temporal structure in ratings and therefore provides an accurate method to predict future ratings. To ensure scalability of ST-KF, we employ a mean-field approximation for inference. Incorporating user/item covariates in estimating neighborhood weights also helps in dealing with both cold-start and warm-start problems seamlessly in a single unified modeling framework; covariates predict factors for new users and items through the neighborhood. We illustrate our method on simulated data, benchmark data and data obtained from a relatively new recommender system application arising in the context of Yahoo! Front Page.
Description
A spatio-temporal approach to collaborative filtering
%0 Conference Paper
%1 lu2009spatiotemporal
%A Lu, Zhengdong
%A Agarwal, Deepak
%A Dhillon, Inderjit S.
%B Proceedings of the third ACM conference on Recommender systems
%C New York, NY, USA
%D 2009
%I ACM
%K cf geolocation item recommendation recommender spatial
%P 13--20
%R 10.1145/1639714.1639719
%T A spatio-temporal approach to collaborative filtering
%U http://doi.acm.org/10.1145/1639714.1639719
%X In this paper, we propose a novel spatio-temporal model for collaborative filtering applications. Our model is based on low-rank matrix factorization that uses a spatio-temporal filtering approach to estimate user and item factors. The spatial component regularizes the factors by exploiting correlation across users and/or items, modeled as a function of some implicit feedback (e.g., who rated what) and/or some side information (e.g., user demographics, browsing history). In particular, we incorporate correlation in factors through a Markov random field prior in a probabilistic framework, whereby the neighborhood weights are functions of user and item covariates. The temporal component ensures that the user/item factors adapt to process changes that occur through time and is implemented in a state space framework with fast estimation through Kalman filtering. Our spatio-temporal filtering (ST-KF hereafter) approach provides a single joint model to simultaneously incorporate both spatial and temporal structure in ratings and therefore provides an accurate method to predict future ratings. To ensure scalability of ST-KF, we employ a mean-field approximation for inference. Incorporating user/item covariates in estimating neighborhood weights also helps in dealing with both cold-start and warm-start problems seamlessly in a single unified modeling framework; covariates predict factors for new users and items through the neighborhood. We illustrate our method on simulated data, benchmark data and data obtained from a relatively new recommender system application arising in the context of Yahoo! Front Page.
%@ 978-1-60558-435-5
@inproceedings{lu2009spatiotemporal,
abstract = {In this paper, we propose a novel spatio-temporal model for collaborative filtering applications. Our model is based on low-rank matrix factorization that uses a spatio-temporal filtering approach to estimate user and item factors. The spatial component regularizes the factors by exploiting correlation across users and/or items, modeled as a function of some implicit feedback (e.g., who rated what) and/or some side information (e.g., user demographics, browsing history). In particular, we incorporate correlation in factors through a Markov random field prior in a probabilistic framework, whereby the neighborhood weights are functions of user and item covariates. The temporal component ensures that the user/item factors adapt to process changes that occur through time and is implemented in a state space framework with fast estimation through Kalman filtering. Our spatio-temporal filtering (ST-KF hereafter) approach provides a single joint model to simultaneously incorporate both spatial and temporal structure in ratings and therefore provides an accurate method to predict future ratings. To ensure scalability of ST-KF, we employ a mean-field approximation for inference. Incorporating user/item covariates in estimating neighborhood weights also helps in dealing with both cold-start and warm-start problems seamlessly in a single unified modeling framework; covariates predict factors for new users and items through the neighborhood. We illustrate our method on simulated data, benchmark data and data obtained from a relatively new recommender system application arising in the context of Yahoo! Front Page.},
acmid = {1639719},
added-at = {2012-09-06T21:53:14.000+0200},
address = {New York, NY, USA},
author = {Lu, Zhengdong and Agarwal, Deepak and Dhillon, Inderjit S.},
biburl = {https://www.bibsonomy.org/bibtex/23c31b37572aac98824304864d4e39364/folke},
booktitle = {Proceedings of the third ACM conference on Recommender systems},
description = {A spatio-temporal approach to collaborative filtering},
doi = {10.1145/1639714.1639719},
interhash = {62268a3ea6e43e1870231bbbb2d6aa37},
intrahash = {3c31b37572aac98824304864d4e39364},
isbn = {978-1-60558-435-5},
keywords = {cf geolocation item recommendation recommender spatial},
location = {New York, New York, USA},
numpages = {8},
pages = {13--20},
publisher = {ACM},
series = {RecSys '09},
timestamp = {2012-09-06T21:53:14.000+0200},
title = {A spatio-temporal approach to collaborative filtering},
url = {http://doi.acm.org/10.1145/1639714.1639719},
year = 2009
}