The Reasreach And Implementation Of Semantic-Based Event Extraction Method For Chinese Text

Event extraction and tracking is a very important area of natural language processing research field. How to extract information of interest accurately and efficiently from large complex and disorder texts has been the key problem of event extraction.In general, event extraction is to extract information of interest from unstructured documents, and at the same time, describe it using a structured form in order to be queried by users and do further tracking analysis. The main study objects should be chosen from some fixed field or news texts, which will match the user s original imagination for event extraction. And the methods are also fixed and monotonous, basically, it will take the pattern matching for structured texts or a paragraph analysis methods, and so on.This subject is based on the research background for semantic search engine with time and geography elements. It proposes a semantic-based information event extraction method through the various innovative combination for rich semantic knowledge and statistical methods. This method emphasizes the tracking function of time and geography elements for event extraction, to implement the description for text event through information choosing and merging.The method will deal with texts with various types,structures and complex drafting style. It will not get the desired results if traditional methods have been used. But this situation is very common in practice. The method proposed by this paper is motivated, effective and not cumbersome. With semantic knowledge and statistical learning, it has a very distinct advantage for complex corpus and large-scale data processing.In addition, the paper includes a lot of NLP concepts and algorithms. It can be said that through this subject, it s easy to have a deeper understanding and insights for NLP, especially in information extraction field.