Modeling item-item similarities for personalized recommendations on Yahoo! front page

DOI10.1214/11-AOAS475MaRDI QIDQ652346zbMATH OpenOpenAlexFDO

Authors Deepak Agarwal, L. Zhang, Rahul Mazumder

Publication date 14 December 2011

Published in The Annals of Applied Statistics (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1111.0416

zbMATH Keywords

matrix factorizationcollaborative filteringrecommender systemsgraphical lassoitem-item similarities

Mathematics Subject Classification ID

Multivariate analysis (62H99) Applications of statistics (62P99) Marketing, advertising (90B60) Communication networks in operations research (90B18)

Abstract: We consider the problem of algorithmically recommending items to users on a Yahoo! front page module. Our approach is based on a novel multilevel hierarchical model that we refer to as a User Profile Model with Graphical Lasso (UPG). The UPG provides a personalized recommendation to users by simultaneously incorporating both user covariates and historical user interactions with items in a model based way. In fact, we build a per-item regression model based on a rich set of user covariates and estimate individual user affinity to items by introducing a latent random vector for each user. The vector random effects are assumed to be drawn from a prior with a precision matrix that measures residual partial associations among items. To ensure better estimates of a precision matrix in high-dimensions, the matrix elements are constrained through a Lasso penalty. Our model is fitted through a penalized-quasi likelihood procedure coupled with a scalable EM algorithm. We employ several computational strategies like multi-threading, conjugate gradients and heavily exploit problem structure to scale our computations in the E-step. For the M-step we take recourse to a scalable variant of the Graphical Lasso algorithm for covariance selection. Through extensive experiments on a new data set obtained from Yahoo! front page and a benchmark data set from a movie recommender application, we show that our UPG model significantly improves performance compared to several state-of-the-art methods in the literature, especially those based on a bilinear random effects model (BIRE). In particular, we show that the gains of UPG are significant compared to BIRE when the number of users is large and the number of items to select from is small. For large item sets and relatively small user sets the results of UPG and BIRE are comparable. The UPG leads to faster model building and produces outputs which are interpretable.

Recommendations

• Smooth neighborhood recommender systems
• Collaborative filtering for massive multinomial data
• An accurate and scalable collaborative recommender
• scientific article; zbMATH DE number 1756678
• Accelerated learning of user profiles

Cites work

• scientific article; zbMATH DE number 3647917 (Why is no real title available?)
• scientific article; zbMATH DE number 4078557 (Why is no real title available?)
• scientific article; zbMATH DE number 3567782 (Why is no real title available?)
• scientific article; zbMATH DE number 1306865 (Why is no real title available?)
• scientific article; zbMATH DE number 739535 (Why is no real title available?)
• scientific article; zbMATH DE number 1049347 (Why is no real title available?)
• scientific article; zbMATH DE number 1134987 (Why is no real title available?)
• scientific article; zbMATH DE number 2107836 (Why is no real title available?)
• scientific article; zbMATH DE number 5485582 (Why is no real title available?)
• Approximate Inference in Generalized Linear Mixed Models
• Arbitrary side observations in bandit problems
• Asymptotically efficient adaptive allocation rules
• Construction of high-order deceptive functions using low-order Walsh coefficients
• Finite-time analysis of the multiarmed bandit problem
• High-dimensional graphs and variable selection with the Lasso
• LIBLINEAR: a library for large linear classification
• Methods of conjugate gradients for solving linear systems
• Mixed-Effects Models in S and S-PLUS
• Model selection through sparse maximum likelihood estimation for multivariate Gaussian or binary data
• One-armed bandit problems with covariates
• Optimal scaling for various Metropolis-Hastings algorithms.
• Randomized allocation with nonparametric estimation for a multi-armed bandit problem with covariates
• Sparse inverse covariance estimation with the graphical lasso
• Trust region Newton method for logistic regression

Cited in

(8)

• Collective matrix completion
• Main effects and interactions in mixed and incomplete data frames
• Smooth neighborhood recommender systems
• Flexible low-rank statistical modeling with missing data and side information
• Collaborative filtering for massive multinomial data
• Two-level monotonic multistage recommender systems
• Bayesian hierarchical rule modeling for predicting medical conditions
• Graph link prediction in computer networks using Poisson matrix factorisation

Describes a project that uses

Uses Software

• S-PLUS
• LIBLINEAR
• glasso
• ElemStatLearn
• MEMSS

This page was built for publication: Modeling item-item similarities for personalized recommendations on Yahoo! front page

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q652346)