A Study Of The Separable Coreference Based On News Events

This paper takes the event extraction task as a perspective,selects the news reports of People’s Daily（2019-2020）for the past three years as research materials,and uses a combination of quantitative and qualitative analysis methods to discuss that the separable coreference has an event triggering function in news events,and analyzes and discusses it in terms of Chinese formal structure and combinatorial semantics,in order to be able to apply this Chinese language feature to Chinese event extraction.The paper is divided into five chapters.Chapter 1 introduces the background of the selected topic and the significance of the study,reviews the status of the separable coreference,event and event extraction and other related research,and explains the research methodology of this paper.The separable coreference studied in this paper is a ternary array of two separations W_A1and W_A2corresponding to a fusion W_F,formalized as<W_F,W_A1,W_A2>.Chapter 2 compares the differences between the separable coreference and the two more similar concepts,the abbreviations and the separable words.Formally,there are one-to-many and many-to-one cases between integrated form and separated form;each character of the integrated form must correspond to one and only one separated form.Semantically,characters of the integrated form can be borrowed directly from the separated form or from characters semantically equivalent to it;the separated forms are semantically independent and divisible.Chapter 3 analyzes the structural types and location distribution of the separable coreference.In terms of structure types,bi-syllabic separable coreference have more structural types than trisyllabic;in terms of location distribution,integrated forms are mostly found in news headlines,while separated form are all found in the text.At the same time,the connection between location distribution and structure type is further analyzed:different distribution locations have a certain tendency to select different structure types.Chapter 4 provides a systematic depiction of the formal structure and semantic constructs of the separable coreference.Taking the integrated form as the base of discussion,four combinations of bi-syllabic integrated forms and multiple combinations of trisyllabic integrated forms are investigated separately.In addition,it is further found that there exists a phenomenon which is the same as the separable coreference in morphological combination but opposite in semantic pointing——the separable cataphoric reference.Chapter 5 examines the rationale for the generation of separable coreference.At the macro level,the universal reason is the driving force of the"economic principle"of language.At the micro level,it is the rhythm of the Chinese language that plays a role;at the same time,it is the information expression of news events.The event-triggering function it assumes can further contribute to the improvement of the event extraction task technology.