Automatic Establishment Of The Hierarchies Of The Dependency Relation Of Chinese Compound Sentence Based On Collocation Of Relation Words

With the society entering the era of big data, every corner of our society is filled with information, and the urgent needs of people for information will promote the rapid development of Chinese information processing. Currently, the Chinese information processing has entered into the phase of sentence and discourse processing after experiencing the phase of word processing and words processing. Chinese compound sentences have abundant and complex semantic information as a bridge that connecting the clauses and chapters. Thus Chinese compound sentence is of important research value in the field of Chinese information processing.Hierarchical identification in the study of the compound sentence is very significance for computational linguistics,syntactic parsing and information project of Compound sentence,but it is difficult to identify hierarchical relationship of compound sentences.Only depend on the hierarchies of compound sentences itself to study the hierarchies of building the compound sentence dependency relation that is difficult to achieve,on the contrary, direct study the hierarchy of building the compound sentence dependency relation will be easy, and it’s the foundation that study automatic identification of hierarchical relationship of compound sentences.As an important component marking the logic semantic relations between clauses in a compound sentence,Relational Markers largely affects the recognition of hierarchical relationship and semantics of clauses.Therefore, this paper presented an automatic constructing method of the hierarchies of the dependency relation of Chinese compound sentence based on collocation of relation words.thus laying the groundwork for automatic identification of hierarchical relationship of compound sentences.In this paper, it is detail analyzed in that how to automatically establishing the hierarchies of dependency relation of compound sentences. Firstly, it introduces the concepts and theories about the hierarchies of dependency relation of compound sentences; secondly, through the analysis of the artificial method for constructing hierarchies of dependency relation of compound sentences, we present the overall work’s step of automatically establishing of the hierarchies of dependency relation of Chinese compound sentence. In this process, the clauses are preliminary division through the dependency syntax and punctuation, and according to the rules to filter out the pseudo clauses. we will use the method to achieve accurate division of clauses in compound sentences.Finally,on the basis of the description of context-free grammar of compound sentences, we present a Algorithm with the shift-reduce based on collocation of relational word to determine the dependency relation between clauses. The syntactic parsing from the language cloud LTP helps to determine the dependency relation between sentence constituents in every clause in order to establish the hierarchies of dependency relation of Chinese compound sentence. Experimental results show that the accuracy rate goes to 84.08% by using the method proposed in this paper. Therefore, it can be considered that the method by present in this paper is effective that automatic constructing the hierarchies of dependency relation of Chinese compound sentence.