Demographic prediction based on user's browsing behavior

摘要: Demographic information plays an important role in personalized web applications. However, it is usually not easy to obtain this kind of personal data such as age and gender. In paper, we made a first approach predict users' gender from their Web browsing behaviors, which the Webpage view treated hidden variable propagate demographic between different users. There are three main steps our approach: First, learning click-though data, Webpages associated with (known) tendency through discriminative model; Second, (unknown) predicted Bayesian framework; Third, based on fact that visited by similar users may be tendency, would visit Webpages, smoothing component employed overcome sparseness log. Experiments conducted real click-through log demonstrate effectiveness proposed approach. The experimental results show algorithm can achieve up 30.4% improvements prediction 50.3% terms macro F1, compared baseline algorithms.