Intranet Browsing Behavior Analysis and Efficient Support System

We introduce a novel approach to analysis of human browsing behavior in electronic spaces, and present an efficient recommender system that helps users increase browsing efficiency by providing suggestions on potentially desired targets of their interactions. Analysis of knowledge worker browsing behavior on a large corporate intranet revealed that users form repetitive elemental and complex browsing patterns. They use relatively narrow spectrum of resources and exhibit diminutive exploratory behavior. Knowledge workers have well defined targets and accomplish their browsing tasks via few subgoals. These findings can be effectively utilized for designing a recommender system that offers users a set of potential targets rather than just the following document in a click stream. It helps increasing browsing efficiency by shortening the required traversal path.