Research On Human Resource Recommendation Algorithm Based On Demand-Aware Collaborative Bayesian Variational Network

With the rapid development of Internet and it technology and the impact of the epidemic situation,the frequency and proportion of enterprises and job seekers to introduce talents and apply for jobs through online recruitment are getting higher and higher,and the human resource information such as the resume of job seekers and enterprise position information on the Internet gradually presents a state of "information overload".For the traditional solution of "information overload",portal website classification and search engine are weak in the case of data explosion and no clear demand of users.Although many recruitment websites integrate classification,search engine and recommendation algorithm to provide information for job seekers,an important longterm problem is how to realize personalized job recommendation considering the future career development needs of job seekers.To solve the above problems,this paper proposes an interpretable personalized human resource recommendation system to promote the application and development of job seekers.The main work of this paper is as follows:(1)after getting the data from the network,the preprocessing operations such as cleaning,conversion and specification are carried out on the obtained data.At the same time,Chinese word segmentation and word bag vector generation are carried out for the fields with more text content,so as to obtain the human resource data for training model.(2)The Demand-Aware Collaborative Bayesian Variational Network(DCBVN)based on demand perception is used for human resource recommendation algorithm.In an interpretable way,the algorithm can jointly model the current ability and career development preference of job seekers.In DCBVN,the potential interpretable representation of job seeker’s ability is extracted from job seeker’s skill files by using automatic coding variational reasoning based on topic model.On this basis,establish an effective demand identification mechanism to understand the personal needs of job seekers’ career development.All the above processes will be integrated into a Bayesian inference view to obtain accurate and interpretable suggestions.(3)Finally,a large number of experiments are carried out with the preprocessed data to show the effectiveness and interpretability of DCBVN in human resource recommendation,as well as its robustness in sparse data and cold start scenarios.