Viewing characteristics based personalized ad streaming in an interactive TV environment

Abstract

The growth in digital TV penetration has resulted in an enormous research initiative in community-specific and home/user-specific target advertising. We propose a channel surfing analysis (CSA) algorithm for predicting the "user(s) in front" of TV by a dynamic analysis of channel viewing characteristics. Such a user identification can be used along with a home information system to decide categorically about the nature of advertisements that are likely to have an impact. The proposed CSA algorithm uses a time-based FSM representation of channel viewing (remote usage) patterns that can be analyzed and processed in multiple ways leading to sophisticated algorithms. Taking into account the fact that not all channel switches provide equal cues about user characteristics, we suggest a multiscale approach to analyze the FSM at various levels of abstraction. We are in the process of integrating the CSA algorithm with an MHP (multimedia home platform) based interactive TV platform that is currently under development.