Research On Database Query Technology Based On Flexible-context Preference

As the e-commerce applications become more and more popular, querying the Web-database is becoming the primary means for obtaining information. The conventional query processing technology simply returns the tuples satisfying the query predicates rather than take users' preference and interest into consideration.In many cases, the user is overloaded with lots of mostly irrelevant information.To resolve this problem, the databse querying techology based on preference has emerged. It considers the users'preference and interest when processing the query. Thus, the results returned can meet users's preferences and make users'query more efficient. Now there are two main approaches to deal with users'preferences, quantitative one and qualitative one. For users' preferences may depend on underlying situations, context-aware preferences start to receive attentions. The existing approaches have some shortcomings:the qualitative one is lack of expressness while the quantitative one can't express the complex preferences. Furthermore, the context in the context-aware preference model is precise. Since the users'query intention is imprecise and not very clear, the precise context can not meet users'requirements.To resolve the above problems, this paper studies the database query techonology based on flexible-context preference. A preference model based on flexible-context and ranking method of tuples returned based on flexible-context preferences are proposed. Adopt the integration of quantitative and qualitative approach to propose a contextual preference model involves interest degree, which has stronger expressive power than each of them and can express the complex preferences. Propose the concept of flexible context and use the fuzzy set theory to extend the precise context, which makes the preference model compliant with users'query habit. Combine all the preferences under the same context and provide the ranked premutation of orders. Then cluster the permutations to generate a few representative ones to save space. When processing users' query, these permutations and their according contexts are used to quickly provide ordered results for users. Exprements and analyses indicate that the preference model based on flexible-context and the ordering method proposed in this paper can meet users' need and preferences. Compared with the same method resolving the problem, the model proposed in this paper has stronger expressing power and has higher ranking quality.