Research On Question Answering System For CNC Machine Tool Fault On Knowledge Graph

With the continuous development of artificial intelligence technology and big data technology,question and answer systems in various fields have developed rapidly.At the same time,the relevant models of constructing knowledge graph are also developing.The question and answer system based on knowledge graph has shown a higher and higher position in various fields.In comparison,unlike traditional search engines,the question and answer system based on knowledge graph will not return a series of matching documents to users like other question and answer systems,allowing users to find their own answers.Instead,it will push accurate answers for users with its intelligent and accurate positioning,which significantly improves the user’s sense of experience and the user’s comfort in using search engine products,It also provides a value orientation for enterprises to solve problems.However,there are relatively few question and answer system models for CNC machine tool faults.How to quickly locate the problem of CNC machine tool faults has become a research hotspot.Therefore,providing users with an accurate and efficient CNC machine tool fault Question Answering System has become an important research project in the industry,and it is necessary to make a knowledge graph Question Answering System related to CNC machine tool faults.At present,the Question Answering System based on Knowledge Graph is mainly divided into two kinds: the method based on semantic classification and the method based on distributed semantic representation.Among them,because the traditional semantic classification methods are easily disturbed by the semantic gap,the recall rate and accuracy are relatively low.With the deepening of research,the way of distributed semantic description is more and more beyond the way based on semantic analysis,but at this stage,the performance of knowledge graph question and answer system method based on semantic analysis needs to be further improved.In previous studies,the accuracy of semantic representation and the relationship between entities limited the research of knowledge graph question answering system.Because these two methods have some problems,this thesis proposes a deep learning knowledge graph question answering system to answer the fault problem of CNC machine tools.The knowledge graph Question Answering System for CNC machine tool fault problems can carry out autonomous learning and accurately answer the questions of maintenance workers because it combines the way of deep learning.Therefore,a knowledge graph Question Answering System model is proposed.This question entity recognition model adds attention mechanism on the basis of ALBERT-Bi LSTM-CRF model,tests the obtained component entity,finds its corresponding triplet,predicts through attribute mapping,finds the candidate answer triplet,and finally reorders the candidate entity triplet through entity link,and integrates the answers through weighted summation.The F1 value of question entity recognition is 6.88%higher than that of the method without adding attention mechanism;and Compared with the model without attention mechanism and string matching,the F1 value of attribute mapping model is increased by 9.67%;The accuracy of the final answer is nearly 2.41% higher than that of the method without adding entity link.