A Study On The Semantic Role Collocation Model Of Mongolian Trivalent Verbs

In the study of Mongolian information processing,verbs are the center of sentences,which can get complete sentence semantics and structure through the semantic roles controlled by verbs.From the perspective of language information processing as a whole,the study of semantic framework is a useful exploration of natural language understanding.The automatic analysis and generation of sentences also need the knowledge of semantic structure.Therefore,this paper aims at the purpose that the machine can correctly recognize the sentence semantics and automatically generate the sentence,annotates the semantic roles of the corpus containing Mongolian trivalent verbs in an exhaustive way,analyzes the characteristics of the semantic roles they dominate,and makes a statistical model of the semantic role collocation of trivalent verbs,and finally creates "the semantic role collocation of Mongolian trivalent verbs" "Matching model information base" and "Mongolian trivalent verb semantic role matching model management software".The thesis is composed of six parts as follows:The introduction part introduces the basis of topic selection,related research overview,research methods and steps,data sources and marker sets,etc.The first chapter introduces the source of the corpus,that is,the selection of words from the "Mongolian verb semantic collocation information database oriented to information processing",the criteria and steps of verb predicate sentence selection from the "one million word level modern Mongolian Corpus",the statistics and sentence processing of words and sentences,the problems encountered in the process of corpus processing and solutions.The second chapter introduces the semantic roles of Mongolian verbs,analyzes the types of semantic roles in the corpus,and makes statistics of semantic roles and semantic role collocation models.The third chapter introduces the construction method and steps of "Mongolian trivalent verb semantic role collocation model information base",as well as the problem of attribute fields and their values.Chapter four introduces the construction method and steps of "Mongolian trivalent verb semantic role collocation model management software"The conclusion summarizes the whole paper,and points out the problems in the research and the work to be done in the future.