Apply incremental classification for a P2P MOD system

With the fast development of technology and science, there come various changes and demands in consumer service market. The plan integrating "broadband network" with "application service" is potential and farsighted in network industry. Academic and technological institutes start researching and promoting the MOD (multimedia on demand) system, which transits multimedia information and provides users interactive interface via setop-box in broadband network. Implement cost, learning content and user behaviors might be taken into account when constructing MOD system. Besides, how to increase the MOD\´s capability of interaction and portability is a practical issue. Sharing and participating are of great importance for network users. This paper first proposes a digital content platform, which combines peer-to-peer (P2P) technology and sharable content object reference model (SCORM). Such a mechanism makes it possible for content providers to share versatile multimedia information with others in digital channel. Followed by a description of file sharing methods in P2P framework. Then, we advocate incremental k-nearest-neighbor (k-NN) algorithm analyzing SCORM metadata and user access records to improve system load, content feature classification efficiency, file index distribution and user query-feedback speed.