Job Recommendation Algorithm Based on Knowledge Graph

Abstract

Aiming at the problem that the traditional collaborative recommendation algorithm ignores the semantic information of post matching personnel in the post recommendation system, this paper proposes a post recommendation algorithm based on a knowledge map and collaborative filtering. Firstly, the knowledge map of employees is constructed by using the data of employees of a company, and the improved knowledge representation algorithm is used to embed the semantic information of employees into a low-dimensional space. Then, the similarity between employees and the score similarity between positions and employees are calculated respectively. Finally, the two similarities are fused and applied to collaborative filtering recommendation. The algorithm in this paper uses the semantic relationship between the recommended objects to alleviate the cold start and data sparsity problems, so that the recommendation results are more accurate. The experimental results on the employee data set of employees of a company show that the proposed algorithm has higher accuracy and recall rate than other traditional recommendation algorithms.