Research And Realization Of Financial Topic Vertical Search Engine

With the development of information science and the dramatic increase of internet information resources search engine has already become one of the most important applications on line. There are many problems among all-pervading search engines in consideration of substantial net resources, such as big-volume information, inaccurate search and shallow depth etc. Therefore, vertical search engine arises. It has the character of profession, accuracy and depth. Furthermore, it can provide valuable information and related services to specific territories and people, as some special demands as well, taking advantage of its industrial feature.This project focuses on the research and realization of vertical search engine on financial topics. The search engine filters out financial information that meets the needs of customers from massive information resource.This article discusses and made research about the relative theories, technologies and realizing processes of vertical search engine, using HERITRIX kit, Lucene fulltext search kit, HtmlPaeser kit and ICTCLAS system to come out the four core modules:webpage grab module, preprocess module, index module and search module. This search engine establishes financial-subjected resource in webpage grab module on basis of vector space algorithm, modifies Heritrix kit to filter hyperlinks. The engine utilizes HtmlParser kit and statistics to slim the web files in preprocess module. In index module the engine modifies Lucene kit to make it supportive for ICTCLAS system and to segment web contents and set reverse indexes which improves the efficiency and accuracy in search. In search module the engine use ICTCLAS and Lucene whole-text search kit, so that the result can meet the needs of users.The lab result shows the vertical search engine does a better job in the satisfaction of customers for financial information.