Research On Information Matching Algorithm In Job Hunting And Recruitment

With the popularity of the Internet,people are more rely on the Internet.The efficiency of publishing and obtaining information is getting higher,which has already changed from the lack of information to the era of great wealth of information.The popularity of the network,makes the job seeker in the face of mass recruitment information,and make the company in the recruitment of facing a lot of resumes.In the job search and recruitment,time-consuming and inefficient.How to pick up the information we need from a amounts of information,realize the job search and recruitment of personalized recommendation.In this paper,based on real data centers of a city as a starting point,extract the related data table about job and recruitment information in the database as the research object.Preprocessing of data table,such as removal of vacancy serious field and repeat entry documents,then word processing,and finally to the match.In a data table,with a text description fields as this article information matching data sets,using the maximum matching word segmentation method for reverse text field,for word segmentation preprocessing and vectorization said.In this paper,using three methods to match.The first,the number of the same vocabulary between job seekers and company vector as similarity,and to sort it.To job seekers as the main body,to recommend 5 individual companies,contrary to the company as the main body,the company recommended 10 candidates.Through access to information related to job search and recruitment,the match is not ideal.The second method,improve the definition of similarity.This is done by means of an external data filtering feature words,give the weight of the words,use of weight calculate similarity,which is the first method of term weight 1 change to the chi-square value,the match is good.The three,similar to the second,also to improve the definite of similarity.The difference is that with its own other properties classified information to filter feature words,give the weight of the words.This method matches the effect is slightly better and closer to the second method.