Research On Causal Event Extraction Of Chinese Financial Texts

Event extraction technology is an important research direction in the field of information extraction.In recent years,with the promotion of evaluation meetings such as automatic content extraction and industrial applications,the research of event extraction has come to a series of theories and algorithms,and put forward some practical systems.At the same time,most of these researches are based on English text extraction,while the research on event extraction of Chinese text is insufficient.Causal event extraction is a branch of event extraction.Compared with event extraction,it is more difficult to extract causal events,and its application field is more detailed,especially in Chinese.In social activities,economic activities play an important role.There are a lot of causal events in the field of Finance and economics.Therefore,according to the previous research on Chinese causal event extraction,it is found that the current causal event extraction methods are mainly based on trigger word template and deep learning methods.The disadvantages of these methods are as follows:First,the extraction methods are based on general Chinese text,thus ignoring the characteristics of the text in the field of Finance and economics.;Second,the current methods based on trigger word template are used The templates are all general templates,and there are few templates;Thirdly,the method based on deep learning needs a large amount of data and poor interpretability.In this paper,a large number of Chinese financial texts are collected from Chinese websites to solve the problem of Chinese causal event extraction in the field of Finance and economics.By refining the fields,we extract events from Chinese Texts in the financial field,and build a Chinese financial event dataset using these events.Based on the trigger word template,this paper decomposes causal event extraction into causal sentence recognition and causal event extraction.By mining a large number of trigger word templates,the causal sentence recognition is optimized.At the same time,Define event entity and event series for meeting the project requirements.At last,a new extraction method is proposed for event extraction.From the perspective of text mining,this paper extracts causal events from Chinese Texts in the field of Finance and economics.In the process of extracting causal events,Optimization of causal sentence recognition by mining causal cues,Proposed a new extraction method for event extraction by combining syntactic analysis and semantic role labeling.