An Aviation Manufacturing Process Knowledge Question-Answering System Based on Knowledge Graph

This paper introduces a question-answering system based on a knowledge graph of the aviation manufacturing process. By collecting and extracting knowledge in the fields of machining, assembly, forming, materials, and other relevant aspects of aviation manufacturing technology, we first construct a comprehensive knowledge graph within the domain. We then leverage natural language processing technologies to address natural language queries. This encompasses named entity recognition utilizing the Aho-Corasick algorithm and Levenshtein Distance, rule-based intention recognition, query template matching and instantiation, among other techniques. Subsequently, answers are retrieved from the established knowledge graph. The test results demonstrate a precision rate of knowledge retrieval at 94.86%, enabling fast and accurate responses to the majority of questions within this domain.