Construction And Application Of Mongolian Word Co-occurrence Network Based On Text Big Data

In nature and people’s lives,the network is ubiquitous.In real life,computer networks,biological networks,and social networks are all representative networks.A network is any system that can be represented by a diagram.This graph is a complex system composed of all the members in a system and some kind of relationship between the members.No matter how large it is,it consists of nodes and edges.With the application of complex network technology in various fields,people regard human language as a complex network system,that is,language complex network.Linguistic complex network is a new branch of quantitative linguistics,which applies complex network technology to linguistic research.At present,there have been many achievements in the research on the complex network of language.Scholars at home and abroad have conducted a lot of discussions on the related issues of English-Chinese word co-occurrence networks,and have been widely used in different language processing.In contrast,the establishment and characteristic analysis of Mongolian word co-existence network is still in its infancy.Exploring the Mongolian word co-occurrence network can fully reflect the overall characteristics of Mongolian,and can be used in Mongolian information processing related work.The main research contents of this paper include:1.Collected and sorted out 10 million word-level Mongolian text corpus mainly on literature,news and law,and completed the processing tasks such as data normalization,text proofreading and root affix segmentation according to the needs of the constructed word co-occurrence network.2.Taking the built text corpus as the data resource,using the graph database development tool Neo4j and the Python programming language,aiming at the characteristics of Mongolian rich in morphological changes,the design and realization of the Mongolian words with the root as the node and the edge reflecting the morphological change a re similar.The current network framework.After the framework is completed,the data is imported,and a multi-word co-occurrence network of directed,undirected,weighted,unweighted,directed weighted,undirected weighted,directed and unweighted,and undirected and unweighted in three different fields is constructed.3.Statistical analysis of the parameters of Mongolian word co-occurrence network nodes,connections(edges),average degree,average shortest path and aggregation coefficient,etc.,and a comparative study of Mongolian word co-occurrence networks in three different fields.At the same time,we also study whether Mongolian word co-occurrence networks have small-world properties and scale-free properties.4.Based on the established Mongolian word co-occurrence network,an information processing-oriented Mongolian word co-occurrence network application-Mongolian associative input method is designed and developed.