An expert recommendation model to electric projects based on KG2E and collaborative filtering

Science and technology projects are an essential starting point for the digital transformation of electric companies. Irrelevant projects' reviewers will cause unrealistic research results. The companies will also waste research funding. We use knowledge graphs and collaborative filtering to recommend experts for projects in the electric power field to solve this problem. First, we constructed an electric project knowledge graph through the project abstract and CNKI database. Then, we use semantic and collaborative similarity to find the experts most relevant to the project. Finally, we discussed the outperformance and conditions of the proposed model and compared the recommendation results with state-of-the-art methods. Research indicates: (1) The knowledge graph can effectively solve the cold-start problem of collaborative filtering. (2) The KG2E-CF model can improve the relevance between the results of the recommendation. (3) The proposed model should be combined with the theme words extraction algorithm to increase the relevance of the recommendation results. Therefore, the expert recommendation in the electric power field can adopt the model proposed in this paper.

Automated summary

This study uses knowledge graphs and collaborative filtering techniques to recommend experts for projects in the electric power field. The research results show that knowledge graphs can effectively solve the cold-start problem in collaborative filtering, and the proposed model can improve the relevance of the recommendation results.