Web Information Search Method Based On Multi-Condition

With the rapid development of Internet technology, Web information has grown explosively. Facing with massive amounts of Web information, the way that users acquiring information usually is submitting keywords or key phrase to the search engine and obtaining information by browsing the returned Web pages one by one. This keyword matching search pattern often meet the search needs when the search query is simple(contains only the keyword or key phrase). However, search engines can't return accurate and comprehensive results when the search query contains many modifiers and complex syntax. The reason for this phenomenon contains multiple aspects. For example, users are not so skilled that they can't accurately describe the search intent for the information they need. More important reasons includes:(1) parts of the search queries contain so much modifiers that the search engines can not accurately resolve users' real query intentions.(2)The final result need to be concluded comprehensively by analyzing multiple Web pages, however, search engines only return a collection of individual pages. In addition, when a user submits a search query contains multiple modifiers, the expected search results always be a collection of entities. By now, users obtain the entity set by browsing the Web pages one by one which wastes a lot of time and energy. To solve above problems, this paper proposes some solutions. The main contribution is as follows:(1)We research the hyponymy between search results and search queries, then analyze, summarize, abstract the Web queries with multiple modifiers. We define the concept model of multi-modifier based Web query. The concept model gives structural description of the query semantic and explains the query process and result set.(2)We propose the strategy of multi-modifier based Web query. According to the strategy, the query is divided into one kernel concept and some modifiers. With the kernel concept and modifiers, we rewrite the query into some sub-queries. We submit the sub-queries to the search engines and get the returned Web pages. We study and conclude how the entities exist in the Web pages and extract entities from Web pages. For each entity, we got its corresponding online encyclopedia pages and compare the attributes in the pages with kernel concept and modifiers. If the attributes in the pages match the kernel concept and modifiers, the corresponding entity will be put in the final result set(an entity set).(3) After studying the differences between the methods in analyzing English search query and Chinese search query, we implement the prototype respectively. As no related multi-modifier based Web search query sets are found, we create the query set both in English and Chinese. We test the function and performance of the prototype with the created query set. The result shows our method is effective.