Trust management for the semantic web

摘要: Though research on the Semantic Web has progressed at a steady pace, its promise yet to be realized. One major difficulty is that, by very nature, large, uncensored system which anyone may contribute. This raises question of how much credence give each source. We cannot expect user know trustworthiness source, nor would we want assign top-down or global credibility values due subjective nature trust. tackle this problem employing web trust, in maintains trusts small number other users. then compose these into trust for all The result our computation not an agglomerate "trustworthiness" user. Instead, receives personalized set trusts, vary widely from person person. define properties combination functions merge such and class merging done locally while maintaining properties. examples specific apply them data Epinions BibServ bibliography server. Experiments confirm that methods are robust noise, do put unreasonable expectations hope will help move closer fulfilling promise.