Research On Vertical Search Method Of University Website Group

With the continuous advancement of the era of online big data,the website information of colleges and universities also continues to grow with the development of the network era.Users need to obtain various college website information through the Internet,and users often search for college website information through a general search engine,but the general search engine It provides users with a horizontal search of massive information in all fields,which prevents users from accurately obtaining the website information they need.In order to solve the accurate search of college website information,a vertical search system for college websites is designed and implemented based on the framework of Solr and Spring Boot.The system consists of five modules: data collection,data preprocessing,text classification,index construction and content search.The data collection module regularly collects teacher information,laboratory information,department information and news updates from the college website group automatically.The data preprocessing module performs data cleaning,text segmentation and word vector training on news texts.The text classification module performs multi-label classification on news dynamic text through convolutional neural networks and long-short-term memory networks.The index building module builds an inverted index of the processed teacher data,laboratory data,department data and news dynamics.The content search module provides users with a search interface,parses user search requests,implements structured aggregate browsing of search results,and improves user experience.Using the school's website group system,the functional test and performance test of the university's website vertical search system were carried out.The test results show that the system has realized the predetermined function,which can realize the accurate search of teachers,academic teams,research directions and other content,and solve the general search.The problem that engine search results are not accurate enough to improve the user's search experience.