The Construction And Application Of Knowledge Graph For Geology

The knowledge graph is a semantic network which contains concepts,entities,events and their relationships in the real world.And it expresses data information like human thinking.At present,the researches of knowledge graphs are mainly focus on the general domains.Construct and present knowledge graph on domain subject have great value.In the field of mineral deposits in geological research,the work of knowledge graph is progressing slowly.This thesis focuses on the construction of high-quality mineral deposit knowledge graph,carry out researches on use of domain knowledge graph such as searching,question-answering.The main researches and contributions of this thesis are as follows:(1)Constructed the deposit knowledge ontology and the deposit entity data set.This thesis uses Scrapy crawler framework and OCR recognition technology to collect structured and unstructured data from multiple sources.Then check and clean the collected data.Create the deposit entity labeling project,carry out the deposit entity labeling work based on the deposit text on the labeling platform,and generate a deposit data set which contains a large number of deposit label entities.Through research on the deposit area,conduct the mineral deposit ontology modeling work,which based on the concept of mineral deposit knowedge classifications and their intrinsic relationships.(2)Propose an improved method of mineral deposit text entity recognition,applied to recognize in many texts,and construct a knowledge graph of mineral deposit.This thesis carried comparative study on classic named entity algorithms,changed the input data and calculation vectors of BiLSTM-CRF model,that is the improved method of mineral deposit entity recognition.By using the improved method,the deposit entities and attributes are extracted from a lot of texts.Storage the mineral deposit knowledge graph data by using Neo4j graph database.(3)The application of mineral deposit knowledge graph.Based on the Neo4j graph database,developed a deposit knowledge graph application system.The system have modules:knowledge search,entity recognition,question and answer,and similarity calculation of deposit.