Human Resource Recommendation Algorithm Based On Latent Factor Model And Deep Forest

With the popularization of the Internet,both individuals and enterprises are more and more inclined to through the network for job search and recruitment.Correspondingly,the number of jobs and user scale are also growing rapidly.Huge user information and job information make the management of human resources information increasingly complicated.Faced with a large number of recruitment information,job seekers can not browse all information due to their limited energy,so that they can't find their interested jobs in time.Traditional solutions,such as information classification and search engine,appear to be powerless when users do not have clear requirements.Therefore,it is particularly necessary to apply the recommendation technology to the human resource field,to excavate the potential interest of the user and recommend the user's interested job.Focusing on the actual application scene,this thesis uses user information and job information and behavior data,and proposed a human resource recommendation algorithm based on latent factor model and deep forest from two angles of user interest and job matching,so as to better perform human resources recommendation.The main work of this thesis is as follows:(1)Data preprocessing and data warehouse: According to the problem of original data,data preprocessing operations such as data cleaning,transformation,and reduction were performed on the collected data.Combining the characteristics of human resources data,the data warehouse design was conducted,and the human resources data warehouse suitable for the proposed algorithm processing was constructed.On this basis,the OLAP analysis was performed.(2)Recommendation algorithm based on latent factor model and deep forest: Firstly,the behavior data was used to construct the biased latent factor model to predict scores and extracted implicit features from the model.Then the implicit features were merged with the explicit features of the user and job into the deep forest for layer-by-layer learning.At the same time,the LR linear classifier was introduced to optimize the cascade forest structure to enhance the classification ability of the model.Finally,combining the user interest degree and job matching degree,the prediction score and the classification result were weightedly mixed to achieve the final job recommendation.(3)Human resources recommendation system: The overall architecture of the system was designed around the target of human resources recommendation,and a human resources recommendation system based on latent factor model and deep forest was implemented to provide personalized human resources recommendation services for users.Finally,this thesis conducted experiments from several aspects to verify the effectiveness of the algorithm.Based on the human resources data warehouse,the traditional collaborative filtering and content-based recommendation algorithms were compared.The experimental results showed that the proposed algorithm has better recommendation performance than the traditional recommendation algorithm in recall and F measure.