Research On The Method Of Fuzzy XML Complex Twig Query With Predicates

Because of its good flexibility and expandability, XML has already been the dominant format and de facto standard for message exchange and expression on the Web. It has been applied widely and deeply to many fields as well. However, the data in the real world is always imprecise and uncertain. At the same time, there is much subjectivity in the real world, such as subjective judgement of an object, and thus some data has fuzziness in nature. It has been an important research topic that how to express this fuzzy information in XML and how to manege these amounts of fuzzy XML documents effectively.As the mainstream structured query language, the core operating object of XQuery is Twig query. Twig query is usually applied to express the nodes and the structural relationship among them which users are interested in. The performance of twig query determines the efficiency of XML query to a great extent. At the same time, there is different logic in user’s query. For example, it must satisfy several different query conditions at the same time(AND logic); it must satisfy at least one of the different query conditions(OR logic); it can’t contain some condition absolutely(NOT logic). At present, there are few research achievements on Fuzzy XML twig query with predicates. On the other hand, it is not perfect on I/O performance for existing twig query algorithms. In view of above conditions, this paper makes further research on Fuzzy XML complex twig query with different predicates.As the existing coding schemes do not support the representation of fuzzy information in Fuzzy XML, this paper expands the extended Dewey encoding at first, called fuzzy extended Dewey. So that it can reveal structure and membership degree information of the node in Fuzzy XML just from the node’s encoding, which lays the foundation for query algorithm. Then, for the reason that the existing query algorithm is not perfect in I/O operation, we propose the algorithms FATJFast, FA/OTJFast and FNTJFast, which settles the AND logic Fuzzy XML twig query, OR logic Fuzzy XML twig query and NOT logic Fuzzy XML twig query respectively. All the new proposed algorithms regard query path as basic matching unit. It is just neccessary to visit leaf nodes instead of all the query nodes during the matching, which reduces the I/O operations directly. On the other hand, they all regard path query as the matching unit, which can cost less time than those algorithms which decompose query twig to finer granularity. At last, we do experiments to test the proposed algorithms. The experiments reveal that the new holistically matching algorithms visit fewer elements than the existing algorithm under the same query condition and improve the general execution efficiency dramatically.