Research On The Evolution Of Ancient Chinese Semantics Based On Statistical Analysis And Distributed Semantic Representation

"A herd of deer sing and eat wormwood in the wilderness",Tu Yo Yo because of artemisinin and renowned at home and abroad,and its success is the help of the ancient Chinese medicine which is derived from the exploration and research.Information technology is increasingly rising,General Secretary Xi Jinping still attaches great importance to and carry forward the Chinese people's most profound soft power,that is,China's outstanding traditional culture.Therefore,the understanding of ancient books,the research of ancient books,not only it is particularly important to inherit and carry forward to the traditional Chinese culture,but also it is of great significance for improving the individual character and life realm.However,the history of Chinese Chinese is long and bright;the evolution of Chinese semantics is complicated and complex,and the understanding of ancient Chinese is becoming more difficult,and semantic evolution is the root cause.The semantic evolution of Chinese is not only an important aspect of language evolution,but also an important research field of historical linguistics.However,in the study of the evolution of the ancient Chinese semantics,only by virtue of the traditional linguistic research methods,can be described as difficult to support.The study of the evolution of semantic evolution in ancient China not only requires the help of science and technology,but also needs the support of the data.And we are both On the basis of the study of traditional linguistics,the combination of statistical(statistical analysis),computational linguistics(distributed semantic representation)and the processing of ancient textual data,the study of ancient Chinese semantic evolution can not only set the three,and can find an unprecedented law,but also create a "calculation of exegesis" to lay a solid foundation.In view of the above analysis,this article uses the ancient Chinese corpus of the ancient Chinese corpus,combined with the relevant research results,specifically carried out the following research:1.Research on the evolution of ancient Chinese semantics based on statistical analysis.First of all,based on the statistical frequency of the research method,that is,because the natural language information is rich in data,complex and diverse features,in the study of Chinese language philology,it is usually used to carry out the qualitative analysis and the quantitative analysis through the method of the statistical analysis,to analyze the specific language phenomenon through the number of relationships,and then reveal its regularity.Secondly,the research method based on Chinese part of speech distribution is used to detect and quantify the evolution of semantic evolution by tracking the changes of the syntactic function of each Chinese character or word.The experimental results show that the research method of ancient Chinese semantic evolution based on statistical frequency has a good effect on the semantic evolution caused by the detection of specific events and the peculiarities of specific entities,and the Research on the Evolution of Semantic Evolution of Ancient Chinese Based on Statistical Chinese Part of Speech Distribution has a good effect on the detection of the evolution of part of speechis of the functional words.2.Research on the Evolution of ancient Chinese Semantics Based on Distributed Semantic Representation.From the perspective of computational linguistics,we use the research method based on distributed semantic representation to study the evolution of ancient Chinese semantics.First,a method of using a distributed semantic representation(the word embedding)based on counting is a method of Explicit Representations(PPMI Matrix)in each sparse matrix of a Chinese character or word in a high dimension and a method of Singular Value Decomposition(SVD)on the basis of the sparse matrix of high dimension,and further investigates the relationship between the deep semantics of each word or word and its semantic evolution.Secondly,based on the method of predictive distributed semantic representation(the word embedding)(based on the Skip-gram model of the Negative Sampling framework)is a method based on SGNS-INC incremental training and a method of SGNS incremental training method.on the basis of improving the training speed and the training quality of the word embedding,these methods can analyzes the deep semantic relation between each word or word and its context in Chinese.The experimental results show that in the study of the semantic evolution of ancient Chinese,we find that the phenomenon of semantic evolution is extremely prominent in the pre-Qin and Qin and Han dynasties,and the phenomenon of semantic evolution in the turn of the Wei and Jin Dynasties and the Sui and Tang dynasties is very prominent.3.Analysis and characterization of the different types of the evolution of ancient Chinese semantic evolution based on distributed semantic representation.The evolution of culture refers to the evolution caused by the change of the natural environment or the social environment.The evolution of the language refers to the evolution caused by the change of the various factors within the language system.Thus,the distinction between cultural evolution and the type of linguistic evolution can be distinguished by detecting the semantic evolution of the representative noun(the evolution of culture caused by the development of technology)and the verb(the evolution of the language caused by the change of the rule).The experimental results show that the global method and the local neighbor method can distinguish the semantic evolution of nouns and verbs,and then make the cultural evolution and language evolution simple distinction.In summary,the main contributions of this paper can be divided into two aspects:1.Interdisciplinary applications.Based on the statistical analysis of ancient Chinese semantic evolution,the natural language information can be quantitatively analyzed,and then its semantic evolution can be detected by frequency jump or part of speech change.The research method of ancient Chinese semantic evolution based on computational linguistics(distributed semantic representation)can make more effective use of its deep semantic information,and then through its evolution in the process of dynasty or period,or the change of similarity in its own detect whether it has a semantic evolution or not.2.Evolution type distinction.According to the different mechanisms of the evolution of culture and the evolution of linguistics,we can distinguish the types of cultural evolution and linguistic evolution by detecting the semantic evolution of their representative nouns and verbs.In general,the study of ancient Chinese semantic evolution based on statistical analysis and distributed semantic representation can co-ordinate the overall situation,and grasp the relationship between the background of the times and the linguistic phenomena presented in ancient Chinese(such as dynasty,The evolution of language and culture).it not only overcomes the traditional linguistic research methods to a certain extent from the perspective of the characteristics of the object,or only limited to the shortcomings of case studies,and broadens the horizon of the development of ancient Chinese semantic new vision,opens up the new ideas in the field of the research of the ancient Chinese linguistics.