Distributed Representations of Web Browsing Sequences for Ad Targeting

Yukihiro Tagami, Hayato Kobayashi, Shingo Ono, Akira Tajima

TargetAd2016(The 2nd International Workshop on Ad Targeting at Scale), 2016/2


Machine Learning Data Science

Large scale user modeling, based on the user activities on the Web, plays a key role in online advertising targeting. In our work-in-progress paper, we introduced an approach that summarizes each sequence of user Web page visits using the Paragraph Vector, considering users and URLs as paragraphs and words, respectively. The learned user representations are used among the user-related prediction tasks in common. In this paper, on the basis of analysis of our Web page visits data, we propose Backward PV-DM, which is a modified version of Paragraph Vector. We show experimental results on two ad-related data sets based on logs from Web services of Yahoo! JAPAN. Our proposed method achieved better results than existing vector models.

Distributed Representations of Web Browsing Sequences for Ad Targeting(PDF 704KB)