Dynamically extracting and exploiting information about customers for knowledge-based interactive TV-commerce

Abstract

Today many e-commerce systems try to extract information about users in order to help users buy products that really need more easily. On the other hand, large TV networks move towards TV-shopping because customers may find it difficult to shop products through the internet due to the lack of familiriaty with computers. In this paper we propose a interactive TV-shopping application, called iTVMobi that extracts and exploits users' information in order to provide automatic personalized reccomendations and adaptive help responses. Our system combines reccomendations and adaptive help to maximise the efficiency of the system towards customer support. iTVMobi has been evaluated by real users. The evaluation shows the the combination of these two functions in an interactive TV environment really helped customers buy products they really needed. © 2008 Springer-Verlag Berlin Heidelberg.