Research And Application Of Wheat Diseases And Insect Pests Question Answering System Based On Knowledge Map

As a major grain crop in China,wheat has a large planting area and high economic benefits,and the occurrence of pests and diseases will directly affect the high quality and high quality of wheat and cause certain economic losses.In this paper,based on the characteristics of multi-source heterogeneous data in wheat pest domain,we use knowledge graph to organize wheat pest domain knowledge,adopt deep learning model to analyze natural language interrogative information,and design and implement a knowledge graph-based wheat pest question and answer system.The main research work of this paper is as follows.(1)Knowledge graph of wheat pests and diseases was constructed.In this paper,the top-down knowledge graph of wheat pests and diseases was constructed by analyzing the data in the field of wheat pests and diseases.Firstly,ten types of entities and eight relationships among entities of wheat pests and diseases are defined,and the conceptual model of wheat pest and disease knowledge graph is designed.In this paper,BIOES is used to sequence annotate the data,and a Bi LSTM-CRF-based method is proposed to identify the information of wheat pest entities.By adjusting the experimental parameters,the F1 value of Bi LSTM-CRF model reaches 88%,which is better in wheat pest entity recognition.Finally,the wheat pest triad data were extracted according to the entity labeling results,and the wheat pest data were stored and displayed using Neo4 j graph database.(2)A BERT-based wheat pest question-and-answer method is proposed.Since there is no publicly available high-quality question and answer text dataset for wheat pests and diseases for the time being,this paper constructs a wheat pest and disease question and answer set by hand.Based on the conceptual model of wheat pest knowledge graph,eight classes of wheat pest question and answer categories were defined.Based on the BERT pre-trained language model,we classify wheat pest question sentences with question intent,and the accuracy of question sentences classification using BERT model reaches 97%.For the entity recognition problem of wheat pest interrogative sentences,the Bi LSTM-CRF model was used to recognize the entities in the interrogative sentences,and the F1 value reached 95.86%.Finally,by forming an interrogative triad with wheat pest interrogative entities and interrogative categories,the user-asked questions are queried in the wheat pest knowledge graph using Cypher statements and presented to the user in the form of natural language.(3)A knowledge graph-based question and answer system for wheat pests and diseases was developed.The question and answer system in this paper mainly uses the knowledge graph of wheat pests and diseases stored in the Neo4 j graph database as the data source,and uses a deep learning model to realize the knowledge question and answer function of wheat pests and diseases,which is an important guidance for wheat pest and disease control.