Optimization Research On Intelligent Search Engine ISMBDI Based Ontology

With the rapid development of network technology and the popularization ofInternet, the WWW is. becoming more and more close to our work and life. It hasbecome a habit to get information from WWW by using search engine. But, with theincrease of information and the manifold of behavior manner, some problems emergedfrom behind the traditional search modes, such as the problem of faithful expressing,and mechanical matching. Essentially, traditional search engine cannot analyze andextend the meaning of keys user inputs and is short of power to handle knowledge and tocomprehend. An intelligent search model based Ontology ISMBDI, uses the theories andtechnologies of Semantic Web and Ontology, improves the information search fromkeyword level to knowledge (or concepts) level, and solves the problem mentioned aboveeffectively. However, ISMBDI has some disadvantages, for example, the user interface isnot friendly enough, the disposal of result can not satisfy users, query efficiency is low.According to the disadvantages of ISMBDI, this paper uses the theories and technologiesof Semantic Web, Ontology and Cache to optimize ISMBDI, so as to improve theperformance and practicality of it and make it become a common share knowledgemodel in the Semantic Web. Concretely, main work in this paper includes several points,such as the optimization of user query interface, algorithms and so on, detailed asfollows.After considering semantic unification and the facility of user interface, a new userquery interfacebased concepts is designed against the problem existed in the user queryinterface of ISMBDI. This UI well integrates the traditional search user interface andthe one of ISMBDI, and not only shortens the time to build a query, lightens the burdenof user, but also guarantees the semantic unification.Based analysis of theprocess of query in ISMBDI, an optimization method of result process is presented. This method solves the problem effectively that there is nothingwhen ISMBDI doesn't have 100％appropriate result, and improves the practicality ofISMBDI. At the same time, in order to implement the method above, a new method ofcomputing the comparability between query and result based on Ontology is presented.This method supplies the comparability between query and result to user by theoperation of the number of schema path mapped successfully and the total number ofschema path contain in query.In order to lighten the burden of worknet and shorten the respond time of ISMBDI,a cache for ISMBDI is built, and correspondingly a new cache replacement policyLRAFU is presented. This policy well integrates the traditional cache replacementpolicies LRU (Least Recently Used) and LFU (Least Frequently Used), not only avoidsthe problem of cache pollution, fits the character of ISMBDI, but also guarantees highhit rate and use rate of cache.According to the optimization method of result and answer the current query byusing, the record in the cache, the key algorithm of query process in ISMBDI-Qplanand Qplan_opt is optimized. The optimized algorithm is called Qplan_imp, it has morefunctions and quicker speed compared with the former.Based on the cache built in ISMBDI, a quick search method is presented. Itshortens the query time greatly and improves the query efficiency, provides a quickpassage for the queries which don't care about the Recall of result, by well using therecord in the cache and losing some result.Finally, the experiments, to test and analysis the key algorithm of query processQplan, Qplan_opt, Qplan_imp, Qplan_quick and cache replacement algorithm LRAFUare implemented. The experiments prove the optimization effect of ISMBDI by using theoptimized methods presented in this paper.