Intelligent Question Answering Technology in Electric Power Field Based on Knowledge Graphs and Improved Editing Distance Fusion

In order to effectively optimize the power business environment, it is essential to build an intelligent question and answer system for business expansion and installation. The intelligent question answering system can efficiently search the relevant content from the massive knowledge base, and directly return the answer to the user, so as to accurately capture the user’s search intention and significantly improve the work efficiency. Therefore, this paper proposes an intelligent question answering system based on knowledge atlas and improved semantic similarity. First of all, the corresponding proprietary thesaurus is constructed according to the actual business requirements, and the triplet data used for training is constructed. Then, the named entity recognition model based on knowledge atlas and the sum of knowledge base are used to extract the entity relationship in customer questions, and the exact answers to the questions are output through triple group matching. Finally, a semantic similarity measurement method with improved editing distance is proposed, which outputs five answers with the greatest relevance to customer questions through fuzzy matching method to improve the reliability of question answering. The performance test of the experiment shows that the proposed intelligent question answering system achieves more than 90% of the question answering accuracy.