Research On Web Databases Fuzzy Relaxation Query Technology

With the rapid expansion of Web databases and data bulk, the database storage objects and methods have been changed. Let user directly access database can be unrealistic. As usual Web databases support a query interface for users to access. But there are more and more users which don’t have enough professional knowledge. When they use the interface to access Web database some problems will happened. For one thing the query intentions of users who can not give accurate value are confused and the query interface that don’t support fuzzy erpressions; for another thing, the user input the explicit query but Web database return empty answers. At this point, normal users may be hope that Web database accept fuzzy expression or return some approximate answers. However, for the current Web database, these requirements are difficult to achieve.To meet these requirements the research works of this paper should solluation three problems:to make Web database query interface support fuzzy expressions; to make interface adaptive to relax the users’ fuzzy queries; to rank query answers and to control the number of query answers. At last the research works give the solution methods and make these into realities. So this paper proposed the strategy of fuzzy relaxation query. The mainly thinking of this strategy is:first, the research of fuzzy query’s relaxation and control. Separating users’fuzzy query into single atom queries then find the set of minimal failing subquery. In each set of minimal failing subquery settle every attribute’s weight. And relax each atom query according to its weight. Combining the deny domain of users’with the membership function given by experts to control the relaxation of atom query. Second, the research of fuzzy relaxation query defuzzification. This part is to reconfirm the edge of attribute value.That need match each attribute value in domain with scope of relaxation. Finally, the research of ranking answers. This paper proposes a computational formula which could rank the query answers by user satisfaction.