Semantic Query Optimization

The technology of Web search has been widely used around the world. However, the precision and recall of existent search engines are not good enough to satisfy user's requirements. At the same time, they are not intelligent enough to carry out some special queries. Presently most search engines are based on keyword or full-text index .As the most important application of the semantic web, semantic search is being paid more and more attention to .The concept of semantic search is put forward in REF. Semantic search integrates the technologies of the search results obtained by current search engines and evolves into the next generation of search engines built on the semantic search..Based on the ontology function in semantic search,we sort semantic search into three types; the incremental semantic search engines based on the traditional search engine, intelligent semantic search based on ontology inference and other semantic search.The characteristics of full-text inspectional is to carry on the comparison to search claim and each phrase in the full-text of customer, taking no account of a matching of search claim and semantic of document, although this kind of way can promise to check whole searching rates, but lower consumedly. What knowledge index emphasizes is according to the knowledge, matching of semantic similarity , so have better assurance in checking a quasi- rate and checking whole rates. The knowledge index is an information index currently the important point of the research, especially face to the knowledge index of Web information. The third generation search engine is trying hard for the customers to use up, provide more convenient and intelligent service. Because all representative search engine at present is according to the search of keyword, however this kind of system usually exist a following problem:(1) In great quantities synonymy and antonym phrases causes to search engine to check whole rates with the existence of many righteousness phrases lower.(2) Usually hand over to imply 1 perhaps 2 the topic related technical term of the customer need only when the customer hands over search to express an information needs, so, in order to handing over search use of the phrase doesn't exert norm and integrity, with text file index usage of phrase or phrase set have very big difference, bringing accurate information index inconvenience.Aim at the above-mentioned weakness of the keyword searching system, people put forward a search an excellent method for turning, that method is with first of search related technical term to join to search at first, become a new search information description. Former of the search is mainly excellent to turn a method to include the method tie of the overall situation analysis a department analysis a method. The basic thought of method of the overall situation analysis is to carry on to the phrase or the phrase sets in all text files related analytical, compute each time to the connection degree of the phrase or phrase set. When a new search arrival, then according to compute in advance of phrase related relation, will join to search with the born and new search at first with the phrase connection degree tallest phrase and phrase set with search. This kind of method can investigate a phrase relation with maximum limit, and carry on search with the higher efficiency after the phrase relate to phrase establishment excellent turn. But, when the text file aggregate very big, building up the relation the phrase of overall situation usually isn't viable in time and space top, and after the text file aggregate change of the renewal price is huge. The part analytical thought is the ex- article which searches first time to think to is a related article, and take this as according to carry on to the search excellent turn. This kind of method is in the current application most extensively, and can use in some actual information index the system. But, at the beginning time search rear row at the text file of front with don't greatly search a related degree at first, the partial analysis will join a great quantities irrelevant phrase into the search, lowering search accuracy thus and seriously, even low don't do excellent turn excellent turn of situation.Making use of to expand an index with righteousness phrase at the inspectional stage can Be not influencing an exaltation under condition of checking the quasi- rate check whole rates, the function raised an information an index system; It is a path which carries out a concept index, the index tool of the network information which has a concept index function can ask for help from 1 together the righteousness phrase watch's keyword to customer importation increases automatically together a righteousness phrase, contribute to an exaltation to check whole rates, but don't lower to check a quasi- rate.There may be various different language expression for each concepts form, can express for the university, college, academy etc. such as the university, search to have something to do with a certain university of concept, the customer inputs probably arbitrarily a key word among them. if input university, match according to the traditional key word of search a method, use college, academy etc. description university of text files all will be left out. use WordNet can know college, academy and university are related, but a little bit other and the irrelevant phrase language of "university", if the jury (jury) also is related with university in the WordNet, if use related concept to carry on searching in brief, although can solve information lapse to a large extent, will bring more useless data in the meantime, make to search to lose meaning.Combine the phrase method connection and the essence language righteousness connection analysis here, excellent turn search process. the customer input several keywords, because all of these keywords is to want to search a contents of elucidation, they may exist a stronger connection very much on the language righteousness. if these keyword and reflect with their phrase method related phrase language to shoot the essence solid example(adoption the key word match of reflecting and shooting method), opposite to say, each other it the language righteousness essence with stronger connection solid example combine, its essence concept is right to represent a customer purpose of the possibility is also bigger.The traditional Web information's index uses keyword as a network information of the search, then lose by customer the person's keyword passes by again a phrase of "with", "or", " not" etc. logic operation of result Be inspectional vector, pass a keyword in the Web page of emergence or not come inspectional. But neglected the inside language righteousness of the phrase an information, chase keyword as a Web information only is the only inspectional population, will bring to include to return to an information necessarily excessive or leak to check an useful information at inside of various problems. So, make the Web information raising the level index from currently make index process according to the language righteousness of the ontology according to the keyword level is matched an evolution by the original keyword set to match for the language righteousness, overcome thus above-mentioned have various blemish that keyword formality bring when matching only, is the root and the key which resolves these problems. This text asks for help from ontology, putting forward according to the ontology of, Web information index prototype system, carried out customer request to match with the contents of network text file to some extent, then raised to check whole rates and check a quasi- rate.Semantic search is different from traditional search. It uses semantic search technology to improve the search results, We develop the semantic search engine. The experiment shows that it can improve recall, through keyword parsing based on the ontology and it can extend keyword to its equivalent concept, sub-concept. The concept query searches all the instances of concept through inference. And the user-defined method can inerrably ensure the semantic information which is hidden in the user's query and improve the precision. Semantic search can also find out the association relationship between two entities. So it can implement some intelligent functions compared to the traditional search engine.The experimentation results show the system can optimize query input efficiently.