User's behavior and URL analysis at EC sites

Abstract

The analysis of user behaviors at EC (Electronic Commerce) sites will allow EC companies to organize efficient web sites and marketers to make effective advertisements for the sites. Such an analysis needs to be implemented on the Internet in order to measure web audiences and ratings related to EC trading. This paper discusses observations made with a user behavior model at EC sites and proposes a new system to measure web audiences at EC sites. The systems use an easy and simple text match to achieve high-speed processing. The user behavior model is proposed based on the characteristics and classification of user behaviors and on URL logs indicating access to EC sites. The user behavior model is used in the proposed system to classify the access logs into four categories: (1) visiting, (2) searching or looking around the site, (3) selecting and ordering items, and (4) confirming the order. More specifically, these categories include user activities at EC sites, such as visiting a web site, searching for particular items, adding some items to a virtual shopping-cart and making order forms on the Internet, and confirming the order request. This URL classification analysis adopts easy and simple text match methods. This paper introduces our prototype and examination.