【摘要】 针对港口机械设备种类繁多，技术繁杂导致信息检索效率低的问题，设计了港口机械智能问答系统。系统采用Protégé编辑本体模型，并依托领域专家的评价构建本体知识库；在此基础上分别基于BERT-BiLSTM-CRF模型和BiLSTM-Attention模型完成命名实体的识别和实体关系的抽取，最后结合SPARQL语句特点将自然语言问句转换为结构化查询语句，检索知识库并输出结果。测试结果表明，该系统能够智能化回答港口机械领域的相关问题，相较于传统的关键词检索方式，检索准确度和效率有极大提升，在工程应用中有一定的现实可行性。更多还原

【Abstract】 Due to the various types of port machinery equipment and its complicated technology,the information retrieval efficiency for port machinery is inefficient.An intelligent question answering system for port machinery was designed.Protégéwas used to edit the ontology model while the evaluation of domain experts was used to establish the ontology knowledge base.The identification of named entities and the extraction of entity relationships for system were completed using BERT-BiLSTM-CRF model and the BiLSTM-Attention model respectively.Finally,the characteristics of SPARQL sentences were combined to convert natural language questions into structured query sentences,retrieved the knowledge base and output the result.The test results show that this system can intelligently answer the related questions in the field of port machinery.Compared with the traditional keyword search method,the retrieval accuracy and efficiency were greatly improved,and it has certain practical feasibility in engineering applications.更多还原