intelliView – An Intelligent Recommender System for Web TV Programs

Abstract

intelliView is a recommender system that suggests TV programs to viewer based on his/her explicit and implicit profiles. This recommender system takes video stream from different Web TV channels and maintains a database of TV programs by scraping program listing from these channels. It manages viewer’s explicit profile by storing programs’ wish list and the implicit profile is created while a viewer watches different types of programs on different timings at different channels. This viewing history in combination with explicit profile is used to recommend programs to the viewer. The recommendation technique used for implementing this system is content-based text categorization strategy. In this technique thousands of feature like words or phrases in the program lists are used to generate programs profile. Later user profile, based on tracking viewing history and explicit wish list, is matched against this programs profile to suggest best match. The technologies used for implementation of the algorithms and user interface are J2EE 2.0, JSF and JDBC 2.0.