Construction Regulations Knowledge Graph Construction And Implementation Of A Question-Answering System

According to the “14th Five-Year Plan”,the construction industry needs to develop a systematic framework,standardize the market,strengthen the engineering safety guarantee system,and comprehensively improve its intelligence level.As the pillar of the global economy,the construction industry needs to update its technology in a timely manner to improve efficiency,which requires strong support from information technology.Currently,China’s construction industry is in a period of high-speed development,and the scale and complexity of the construction market are increasing day by day.However,this also leads to a lack of fast,accurate,and comprehensive ways for people to obtain construction regulatory knowledge.In order to solve this problem,this thesis relies on data such as construction laws and regulations to conduct research on knowledge graphs and designs and implements a question-answering system based on the knowledge graph of construction laws and regulations.The system is mainly aimed at personnel studying and taking exams on regulations in the construction field and construction practitioners,aiming to provide a tool for managing and visualizing knowledge.The knowledge graph of construction laws and regulations,as an industry knowledge graph,has not yet matured into a construction laws and regulations knowledge base in the current research,so the construction laws and regulations question-answering system proposed in this thesis has strong application value and practical significance.The main contents of this thesis are as follows:(1)From the four aspects of bottlenecks in the engineering field,problems in the construction market,lack of knowledge acquisition channels,and policy support for the construction industry,this thesis introduces the research background in detail,and introduces the research purpose and significance.The progress of knowledge graphs and question-answering systems is comprehensively expounded,and the research status of knowledge graphs in the construction field is analyzed.(2)This thesis explores the construction technology of the knowledge graph of construction laws and regulations,including data preprocessing,information extraction,and knowledge fusion technologies.Through these technologies,the construction laws and regulations text data are converted into structured data,and Neo4j graph database is used for storage and management to successfully build the knowledge graph of construction laws and regulations.In terms of named entity recognition,this thesis uses the manually annotated BIO construction regulations corpus as the training and testing dataset,and constructs a building regulations named entity recognition model based on ALBERT+BiLSTM+CRF.Through comparative experimental analysis with other models,the rationality and advantages of this model are verified.(3)Based on the knowledge graph,this thesis constructs a knowledge question-answering module.The entire module can be divided into three sub-modules: question parsing,knowledge retrieval,and answer ranking.In the question parsing module,various techniques such as the Aho-Corasick algorithm,named entity recognition,and TextRank algorithm are used to parse natural language questions.In the knowledge retrieval module,Cypher statements are constructed based on the information parsed from the question to efficiently retrieve information from the knowledge graph and obtain the answer set for the question.Finally,by calculating the similarity between the answer set and the question and sorting it,the best question answer is returned.These modules collaborate with each other to ensure that users can easily ask natural language questions and quickly obtain accurate answers.(4)Based on requirements analysis and overall design,this thesis has developed a question-answering system for an architectural regulations knowledge graph,which includes functional modules such as building regulations Q&A and knowledge visualization.The system aims to meet users’ needs for architectural regulations-related questions and provide them with more convenient and efficient knowledge services.The system adopts Django and Bootstrap as the front-end and back-end frameworks,ensuring scalability and maintainability.Additionally,Neo4j graph database is utilized for storing and querying the knowledge graph,enabling efficient knowledge management and search functionality.Furthermore,the Echarts tool is employed for visual representation,allowing users to gain a more intuitive understanding of the information within the knowledge graph.The construction of a knowledge graph and a question-answering system for building regulations is a preliminary attempt and application of knowledge graphs in this field.It combines building regulations with artificial intelligence,forming a highly practical and scalable system.The establishment of this system not only provides a rich knowledge base for subsequent researchers,but also offers a high-quality and standardized learning and reference platform for professionals and students in the building industry.