Research And Implementation Of The Domain-Dependent Vertical Search System

With the crazy popularity of Web2.0, the expansion speed of network information resources is also growing exponentially, massive data resources far beyond the scope of the search engine can cover, in order to use the traditional search engine technology quickly and accurately find the required information becomes increasingly difficult. Alarming rate of increase of the amount of data makes the general search engine is difficult to promptly update the index database; vast web resources to make the general search engines crawl-depth information becomes more difficult. In response to these shortcomings, a new generation of search technology-vertical search engines came into being.Vertical search engine is the breakdown of general search engines and extension, only for a particular industry or subject, for the specific group of people with valuable information and related services. Topics crawling and retrieval services as two important aspects to a large extent affect the vertical search engine query accuracy and efficiency of retrieval. How quickly and efficiently determine and predict the topic pages and how to give the user clear and accurate feedback as a constraint to retrieve the vertical search engine development the two problems. Therefore, how to improve these two aspects becomes the starting point of this article.In this paper, the directory-type classification based on the theme description, and thus subject to achieve a new crawling strategy. Focused crawler makes relevant to the subject no longer blindly predict the direction of the page, but by giving the location of the ODP in the subject of different levels of different node weights more accurately guide the crawling reptile theme. Terms in the search results presented in this paper the advantages of using clustering search engine, phrase-based topic approach, feature items on the document a more accurate extraction and clustering to guide the work, presented with a cluster to give users a more convenient way query experience. Finally, this paper verify the validity of the two through the design of comparative test to verify the validity of the two.