Exploiting content relevance and social relevance for personalized ad recommendation on internet TV
ACM Transactions on Multimedia Computing, Communications, and Applications (TOMCCAP)
A cloud-based intelligent TV program recommendation system
Computers and Electrical Engineering
| Hi-index | 0.43 |
This paper presents the novel development of an embedded system that aims at digital TV content recommendation based on descriptive metadata collected from versatile sources. The described system comprises a user profiling subsystem identifying user preferences and a user agent subsystem performing content rating. TV content items are ranked using a combined multimodal approach integrating classification-based and keyword-based similarity predictions so that a user is presented with a limited subset of relevant content. Observable user behaviors are discussed as instrumental in user profiling and a formula is provided for implicitly estimating the degree of user appreciation of content. A new relation-based similarity measure is suggested to improve categorized content rating precision. Experimental results show that our system can recommend desired content to users with significant amount of accuracy.