Knowledge Map Construction And Value Analysis Of The Proper Name Of "The Classic Of Mountains"

In about 20000 words,Shan Jing introduces a large number of mountains,valleys,plants,birds and animals with a rather structured narrative.Its structural and multiproper nouns/names narrative features make it very suitable to be used to present knowledge graph.But at present,thers is no scholar having such a try.Using knowledge graph technology to visualize the proper nouns/names and their relationships in book Shan Jing,the main tasks involved include: Determining and classifying the proper nouns/names in Shan Jing,and constructing the knowledge graph of the proper nouns/names and their relationships.The latter can be subdivided into the following steps: obtaining and proofreading the electronic text of Shan Jing,constructing entity dictionary,extracting entities and their relationships by using regular expressions,entity disambiguation and entity alignment,drawing knowledge graph in Neo4 J graphic database.When determining proper nouns/names,we mainly define them based on the theories of " hierarchical subdivision" and "mutual transformation" put forward by scholars.And then count and classify them into 14 categories: Mountain name,water name,plant name,tree name,bird name,insect name,mineral name,aquatic life name,God man name,valley name,etc.A total of 1341 proper nouns/names were counted,among which mountain names were the most,while shellfish,turtles and other aquatic names were less.Because some proper nouns/names have the phenomenon of "the same name with different reference" or "the same reference with different name",it is necessary to analyze these proper nouns/names by consulting relevant literature.The construction of knowledge graph is mainly based on the principle of authenticity and integrity.In order to present the knowledge graph of Shan Jing as faithfully and completely as possible,some special entity cases that can not be called proper nouns/names will be added in the entity dictionaries,and some entities that several experts think "may be the same reference" will not be aligned during th e process of entity alignment.Because the text of Shan Jing is very structured in narration,entities are mostly connected by fixed verbs or verb phrases.Therefore,when designing regular expressions template based on verbs or verb phrases for entity extraction,relationship extraction is also completed synchronously.In this paper,the extracted entities and entity relationships are saved by adding category or number labels the entities and defining storage templates.It should be pointed out that adding category labels to entities is also one of the means to implement entity disambiguation in this paper.The article has completed entity and relation extraction with the accuracy rate and recall rate higher than 95%,so the knowledge graph drawn on this basis has retold the Shan Jing text to a large extent.Reviewing the experimental process of this paper and thinking about the possible application scenarios of the experimental results,we can see the value of this experiment from the following two aspects: First,the value of using regular expressions to study the proper nouns/names of Shan Jing;Second,the value of the knowledge graph in the text query and digitization of ancient books.The first value is mainly for the textual criticism of Shan Jing.Using regular expressions to study the proper nouns/names of Shan Jing helps to find the text errors in Shan Jing and promote the unification of punctuation system in Shan Jing;The latter value is related to the digitization of ancient books.The knowledge graph generated in this paper can help scholars to query the text of Shan Jing and find the hidden information in the text,and further build the knowledge base of ancient books on this basis.