抄録
World Wide Web can be said to be the most prominent media of information offering and obtaining. In our time the start point for acquiring information from the web is seeking information using the search engines that direct us to many possibly relevant web pages. The user is responsible then for estimating trustworthiness of a page on the subject and combining the information form different pages to arrive at a unified unique knowledge about the subject. The semantic web envisions a time when machines do such jobs as trust estimation and information composition. This work proposes a combining procedure for information obtained from different sources according to estimated trust to the sources. The trust to sources plays an important role in integration of information from them. According to the vague nature of human language statements and trust to source estimation, an extended fuzzy description logic where a membership value and a numeric type individual can also be a fuzzy set is suggested to represent the formal representation of assertions. Two fuzzy rules are proposed for combining fuzzy description logic assertions and are illustrated in action through a practical example. It is shown that the composition procedure performs well at integrating even contradictory information from different sources by taking the trust to sources into account.
