Research on Knowledge Driven Intelligent Question Answering System for Electric Power Customer Service

摘要

The electric power customer service domain-specific knowledge graph aims to describe the concepts, entities, events and their relationships in the electric power customer service business field, with a structured manner, and provide a more effective data organization, management, and cognitive ability for the customer service business. This paper proposes a method for constructing a knowledge graph in the field of electric customer service, and constructs a knowledge graph with over 15,000 entities and 20,000 relationships. Based on the knowledge graph, an intelligent question answering application architecture is designed, which consists of multiple functional modules such as dialogue process configuration, natural language processing, and business process processing. It provides more efficient and open knowledge retrieval services for the electric customer service business, and improves the intelligence level of customer service question answering.