Research And Implementation Of Vertical Search Engine On English Sentences

The Internet is the main tool of the exchange of information and plays an important role in the modern society. During the period of the beginning of Internet, the amount of information is not large. However, with the rapid development of Web technology, the Internet has a huge number of pages. Search engine helps people to retrieve the effective data from the large web quickly and easily. However, traditional search engines are unable to fit the user’s precise search on the subject information. And vertical search engine can solve such problems. Meanwhile, the vertical search engine has become the development direction of the search engine.This paper first discusses the research background and meaning of traditional search engines and vertical search engine. And we analyze two types of search engine, such as system structure, key technology, the working process and the difference. And then this paper implements the system of English sentences vertical search engine based on the enterprise search engine Solr. In order to identify the authority of the English sentence conveniently, we research the search ordering results deeply. The aim is to find an English sentence scoring model to show authentic English sentences instead of manual work. This experiment first analyzes and extracts text features such as words, phrases, sentences. According to correlation between the user’s real score and text features, the relevant feature set is extracted. Then using feature set to do Principal Component Analysis is to reduce the number of feature items. And then five representative features is selected to do Regression. In the end, building up a reasonable model for English sentences is to score and recommend authentic English sentences to user.Research and Implementation of the subject constitutes vertical search engine system on English sentences. As a core module in the subject, the ordering strategy for English sentences is new for search engine optimization. This method can be reasonably optimized to the site operation.