Exploiting User Preference for Online Learning in Web Content Optimization Systems

Abstract

Web portal services have become an important medium to deliver digital content, e.g. news, advertisements, etc., to Web users in a timely fashion. To attract more users to various content modules on the Web portal, it is necessary to design a recommender system that can effectively achieve Web portal content optimization by automatically estimating content items' attractiveness and relevance to users' interests. The state-of-the-art online learning methodology adapts dedicated pointwise models to independently estimate the attractiveness score for each candidate content item. Although such pointwise models can be easily adapted for online recommendation, there still remain a few critical problems. First, this pointwise methodology fails to use invaluable user preferences between content items. Moreover, the performance of pointwise models decreases drastically when facing the problem of sparse learning samples. To address these problems, we propose to explore a new dynamic pairwise learning methodology for Web portal content optimization, in which we exploit dynamic user preferences extracted based on users' actions on portal services to compute the attractiveness scores of content items. In this paper, we introduce two specific pairwise learning algorithms, a straightforward graph-based algorithm and a formalized Bayesian modeling one. Experiments on large-scale data from a commercial Web portal demonstrate the significant improvement of pairwise methodologies over the baseline pointwise models. Further analysis illustrates that our new pairwise learning approaches can benefit personalized recommendation more than pointwise models, since the data sparsity is more critical for personalized content optimization.

Timeliness of Materials on Reading Recommendation System

“…The most common methods are Collaborative Filtering algorithm, Content-Based algorithm and Hybrid algorithm [3]- [5]. These methods are successful in the practice of recommendation system, but they still remain some problems, such as cold start problem, data-sparse and some important factors of reading materials are not be effectively applied [6].…”

Temporal dynamics in recommender systems

“…Pairwise Preference-based Method The methodology of dynamic pairwise learning can be applied to improve the performance of RSs [34]. The pairwise learning method is developed based on the implicit feedback extracted from users' actions on portal services.…”

“…A probabilistic model based on Bayesian hidden score method [222] is developed to generate user segment-based ranking and recommendations [34]. Let r i,j;c t denote the perceived preference between content i and content j for user segment c at time interval t. The…”