Discovering Behavioral Traversal Patterns of Users of Streaming Information Systems

摘要

Several research efforts have been directed towards mining the navigational patterns of users of information systems. Previous research on identifying navigational patterns was directed at browsing patterns of web users. The contents of web pages do not change frequently, thus, existing approaches for discovering navigational patterns are guided by this philosophy. In this work we motivate a new domain for data mining, which involves mining the navigational patterns of users in environments with streaming content (for example, cable-TV viewers). The access sites we study in our framework display continuously changing content unlike web pages that are relatively static. This fundamental difference in site characteristics calls for new algorithms for identifying interesting navigational patterns. The algorithms proposed in this work are generalized and can be applied to any system with similar characteristics. The approach used in this paper is based on behavioral predicates, and incorporates the rich temporal semantics existing in sites with streaming content.