A Method for Supplementing the Knowledge Graph of Power User Service

Abstract

In order to reduce the workload of manual supplementation of the knowledge graph of power user service and improve the efficiency of supplementation, a method that can assist manual supplementation of the knowledge graph of power user service is designed. This paper firstly introduces the structure of the knowledge graph of power user service and the design scheme of the graph supplementing method based on human-computer interaction, and defines the questions that need add into graph and the process that how to supplement the knowledge graph. Then, the key technology in the graph supplementing method—user questions matching, is introduced. A deep learning-based user questions matching method is proposed and an algorithm suitable for questions matching in the power user service is selected through experimental comparison. Finally, analysis and summary are carried out.