Font Size: a A A

The Design And Implementation Of Chinese Knowledge Graph Ontology Building System

Posted on:2020-10-25Degree:MasterType:Thesis
Country:ChinaCandidate:H T QinFull Text:PDF
GTID:2428330590959830Subject:Software engineering
Abstract/Summary:PDF Full Text Request
In recent years,the research of information extraction,knowledge map and the construction of knowledge base has become a hot topic.Entity recognition and entity lower relation extraction are the basis of these research.However,there are not many ontology constructions for specific knowledge maps at present.This system has done some work for ontology construction of knowledge maps in plant field.This system mainly uses a bottom-up way of building,starting from the bottom of the domain entity,mining epistasis through the recommendation algorithm,avoiding the difficulty of the relationship extraction and relation judgment in the traditional construction method,obtaining the entity epistasis through the automatic way,and then getting the epistasis of the epistatic word.Hypernym hierarchy,finally through information extraction and mining all the attributes of the entity,finally constructed ontology.First,the acquisition of domain entities,the system through mining the Internet open database data,such as Sogou,Baidu,Tencent and other Internet Co collection and collation of various fields of words,based on such terms to enrich the domain entity base.The second step is to crawl the encyclopedia data,obtain the epistasis of the entities collected in front of the encyclopedia and open classification,and clean the collation.In addition,the epistasis is obtained by using the cooperative transition recommendation algorithm as a supplement to complete the epistasis acquisition.In the third step,the method of searching frequent items by FP-Growth algorithm is used to level the epistatic words obtained in front,and the epistasis of the epistatic word is found until the end of the epistasis.Finally,the text of the encyclopedia described,information extraction,mining the related properties of the entity,complete the whole process of the construction of the whole ontology,and displayed in the form of atlas.This paper focuses on the above four processes,through systematic analysis,summed up the five main function modules of the system,then design and implement.In the whole process of design and implementation of the system,we have a deep understanding of encyclopedia knowledge crawling and algorithm implementation.In this paper,we do some research on theoretical analysis and practical application,and design and implement the plant domain ontology construction system.
Keywords/Search Tags:knowledge map, ontology, superordinate word, hierarchy, attribute extraction
PDF Full Text Request
Related items