The Research Of Fuzzy Query Results Ranking Based On Relational Database And Its Application In Realestate Information Search System

In real applications,the information is usually vague or imprecise, thus people would like to express their selection query by using flexible query on relational databases, and such queries are called fuzzy queries. However, the fuzzy query may result in too many answers over a large database, and the users are only interested in the results which meet their intention most closely. So it is important to rank the query results of the fuzzy queries over relational databases.This article proposes an automated ranking approach for the fuzzy query results for database queries. Our solution leverages data and workload statistics to capture the user preferences, and so it can formulate a ranking requirement in order to return the ranked fuzzy query results for the users. The ranking function this article proposes measures the importance of the tuple by regarding its satisfaction degree of specified attribute values to the fuzzy query, and its relevance degree of unspecified attribute values to the user's preferences. Firstly, the tuples are ranked according to the membership degrees of specified values of answer tuples to the fuzzy query. Secondly, based on the database workload and data statistics, an estimation method of relevance degree of the unspecified attribute values of answer tuples to the user preferences is proposed, which is used to sort the tuples sharing the same membership degree. And finally, we rank the query results according to the satisfaction degree of the specified attribute values to the fuzzy query and the relevance degree of the unspecified attribute values to the user preferences.The proposed approach for ranking the many fuzzy answers is finally used in a real estate information search system. During the real application, the approach for ranking the many fuzzy answers from the relational database proposed in the article can capture the user's preferences effectively and have a high ranking quality as well, which demonstrate the correctness and feasibility of our ranking approach.