Algorithm Design And Implementation Of Personalized Job Recommendation System Based On Web Information

With the continuous development of computer networks,online job search system has gradually become the mainstream way of job search.At the same time,the recruitment information has exploded.The application of personalized work recommendation can help job seekers find jobs that meet their needs faster.The existing job hunting system is mainly centered on enterprises,and there is false recruitment information.Secondly,most job recommendation systems mainly use resume information from job seekers to recommend.Once the information security problem arises,it will bring incalculable loss to job seekers.In addition,inappropriate recommendation or too low recommendation frequency will bring a certain degree of resentment to job seekers.On the basis of in-depth study of the existing recommendation system related algorithms,this paper designs a personalized recommendation algorithm(J_Rec)based on Web user behavior information,combining the personality of the work recommendation and the actual requirements of the job search process.In this paper,the algorithm design is divided into four parts:recommendation user selection,user behavior analysis,hybrid recommendation algorithm and generating recommendation results.First of all,we recommend user selection,combined with the timeliness of job search needs,to determine whether users do have job requirements through online time.Secondly,the user's Web log behavior data is used to analyze the potential needs of users,avoid users' personal information,and protect users'privacy.Third,combined with popular recommendation,association recommendation and personalized recommendation,from the perspective of the system and users,it is effective to deal with cold start and better meet the user's different job requirements.Fourth,the proposed results take full account of the timeliness of job and job search requirements,as well as the feasibility of off-line recommendation,and recommend more suitable jobs to users.The algorithm experiments are carried out through data set Lagou-data(data obtained from crawler technology from "Lagou net").It also contrasts four indexes of precision,recall,coverage and FI-Measure with user based collaborative filtering(User-CF)and project based collaborative filtering(Item-CF)algorithm.At the same time,the similarity between User-CF and Item-CF is compared with MAE performance of Cosine similarity,Pearson similarity and Jaccard similarity.The experimental results show that the Jaccard similarity formula is the best.Compared to the three algorithms,J_Rec is superior to User-CF and Item-CF on precision,recall and F1-Measure.With the increase of the number of recommended users on coverage,the final results of the three algorithms tend to be consistent.