Design And Implementation Of Intelligent Question Answering System In Coal Mine Industry

Nowadays,people are accustomed to obtaining information through Internet search engines such as Baidu and Google,but the web page data returned by the Internet is usually messy.Users need to spend time and energy to judge the redundant information,and they cannot directly get accurate answers.The knowledge of the coal industry is distributed and hashed,and it is difficult to find quickly.The intelligent question answering system based on the knowledge graph can capture the user's intentions by analyzing the questions asked by the user using natural language,and search the knowledge base to find the most accurate answer directly to the user.The main research contents of this topic include:(1)Knowledge base construction: use Scrapy to crawl and download the files on the coal mine knowledge website to obtain data sources,and then convert the file format,and then preprocess the file content through LTP,and use the word-based Bi LSTM-CRF to name entities The recognition model and the relation extraction model based on Bi GRU fusion Attention mechanism obtain the structured data of <entity 1,relation,entity 2> from the data in the marked text,and store it in the Neo4 j graph database to form the coal mine knowledge base.(2)Question and answer implementation: first crawl the common question sentences,obtain the question types frequently asked by users,construct the question sentence template,then preprocess the question sentence set,use TF-IDF to extract the characteristic words of the question sentence,and form the characteristic word dictionary.Train the question set using support vector machines and naive Bayes respectively to obtain question classifiers,test and analyze the training results of the two classifiers,and finally select the naive Bayes classifier with better effect for users The question sentence entered in the question and answer system is classified,matched with the question template,and the complete Cypher query sentence is obtained.Then,the answer is searched in the coal mine knowledge base,and the answer is filled into the answer template and returned to the user.(3)Front-end interaction of the system: This article uses the Django framework to realize the connection and data interaction between the question and answer system and the coal mine knowledge base,and uses Bootstrap and JS to realize the front-end interface,and finally a complete question and answer system is obtained.The research of this subject can improve the efficiency of obtaining information for professionals in the coal mine field and reduce the waste of time.It also has important reference value for the research of knowledge base question and answer in other fields.