基于最大熵的越南语新闻事件元素抽取方法

越南与中国一水相依，是重要的政治、军事和经济合作邻国，然而针对越南语新闻事件元素的提取研究非常匮乏。本文针对越南语特点，提出一种基于最大熵模型的越南语新闻事件元素抽取方法。该方法针对越语句子结构和词汇语义的特点，采用最大熵算法，选取上下文、邻近触发词以及邻近实体作为特征，定义特征模版，训练获得越南语新闻事件模型，实现新闻事件元素抽取。抽取实验结果表明本文提出的方法抽取新闻事件元素的准确率达到80%以上。

Extractiond Method of Vietnamese News Event Elements Based on Maximum Entropy

The study on extraction of Vietnamese news event elements is rare, while Vietnam is a significant neighboring country with political, military and economic cooperation, which is just at a distance of a river with us. According to the Vietnamese characteristics, this paper puts forward a method of Vietnamese news event element extraction based on maximum entropy model. This method selects the context, adjacent trigger words and neighboring entities as features, delimits feature templates, trains Vietnamese news events model and achieves the extraction of news event elements of Vietnamese on the basis of the characteristics of the Vietnamese sentence structure and lexical semantic using the maximum entropy algorithm. The experimental result of the extraction shows that the accuracy of the news event elements extracted by the method proposed in this paper reaches more than 80%.