Parallel neural networks for multimodal video genre classification
Multimedia Tools and Applications
Hi-index | 0.00 |
In order to achieve TV program group recommendation,an approach based on fuzzy clustering is proposed forprogram classification in this paper. This paper firstlydescribes the XML based program description metadatarepresentation, in which both textual and symbolicinformation is included. Secondly it presents the programfeature extraction and presentation method. A program isdefined as two vectors, one is based on term statisticsimplying what the program is about, and the other reflectsbroadcasting characteristics of the program. Then theclassifying approach based on fuzzy clustering is proposed.The approach goes: normalizing original data, buildingfuzzy similarity matrix, and then clustering. The final fuzzysimilarity matrix is constructed by combining two fuzzysimilarity matrices calculated from two different aspects.