Kiwifruit Planting Knowledge Question And Answer System Based On Knowledge Graph

In recent years,the kiwifruit planting industry in China has developed rapidly,and according to the FAO 2019 data,its planting area and annual production are in the 1st place in the world.The yield and quality of kiwifruit are affected by factors such as origin,climate,planting technology and pest control,but the existing kiwifruit planting and management knowledge is stored in the form of traditional database,which has low knowledge utilization rate when information is queried,and the search engine approach cannot give accurate answers according to user questions,and the user experience is poor.In contrast,knowledge graphs store knowledge in a structured form and can better correlate knowledge to maximize information utility.The design and implementation of a knowledge graph-based kiwifruit planting knowledge Q＆A system can provide efficient and accurate Q＆A services for planters.In this study,the research is carried out in three aspects: the construction of kiwifruit planting knowledge graph,the method of kiwifruit planting knowledge Q＆A based on the knowledge graph,and the design and implementation of kiwifruit planting knowledge Q＆A system based on the knowledge graph.The main work and results of this study are as follows.（1）Construction of kiwifruit planting knowledge graph.The top-down approach is adopted to construct the knowledge graph of kiwifruit planting based on the characteristics of multiple sources of heterogeneous data in the field of kiwifruit cultivation,and the ontology concept model of the knowledge graph of kiwifruit planting is designed first,and then the extracted triadic facts are added to the knowledge base of the data layer according to the ontology specification of the model layer.A joint entity relationship extraction method and a TransR-based knowledge complementation method are used,and a kiwifruit planting entity recognition model incorporating word-meaning information is constructed.The model is based on SoftLexicon,and the accuracy of named entity recognition is further improved by adjusting word weights and word set importance through MHA and Attention mechanisms,respectively.The experimental results show that the F1 value of the kiwifruit planting entity recognition model constructed in this paper improves by 1.58% to 91.91%compared with the SoftLexicon model,and the F1 value reaches 96.17% on the ResumeNER public dataset;the F1 value of kiwifruit planting triad extraction is 92.86%;the knowledge complementation methods based on TransR Hit@3 and Hit@10 were 90.40% and 92.60%,respectively.（2）Research on the knowledge question and answer method for kiwifruit planting based on knowledge graph.For the problem of complex structure,various expression forms and uneven quality of user natural language question and answer,the heuristic extension-based subject entity matching method and BERT-based question intention classification method are proposed to understand user natural language question and answer;for the problem that user natural language question and answer is difficult to link to the knowledge graph correctly,the BERT-CNN-Siamese based The relational chain linking method is proposed to link user question sentences with relational chains;by adding entity constraints to relational chains,the accuracy of kiwifruit planting knowledge quiz is further improved.The experimental results show that the accuracy rate of the topic entity matching method is91.32%,the F1 values of the question sentence intention classification method and the relational chain linking method are 95.07% and 95.48%,respectively,and the answer query accuracy rate is 82.86%.（3）Design and implementation of a knowledge-based Q＆A system for kiwifruit planting based on knowledge graph.In view of the fact that the existing Q＆A system does not realize the knowledge Q＆A service of kiwifruit planting based on the knowledge graph,the design and implementation of the knowledge Q＆A system of kiwifruit planting based on the knowledge graph is studied.This study takes the kiwifruit planting knowledge graph as the data base,Django and ECharts as the development framework and data visualization tools,and the kiwifruit planting knowledge Q＆A method based on the knowledge graph as the business core,and builds a knowledge graph-based kiwifruit planting knowledge Q＆A system to provide a kiwifruit planting knowledge Q＆A service platform for planters and promote the intelligent development of the kiwifruit industry.