自然语言文本中不确定性信息的自动识别

自然语言中存在大量不确定的表述,针对此类信息的检测任务是信息抽取领域的研究热点之一,然而,面向中文的不确定信息检测研究仍然比较匮乏,利用支持向量机(Support Vector Machine,SVM)能够很好的解决非线性、高维数、局部小样本等实际问题的优势,将中文不确定性信息识别问题转化为分类问题,通过在复旦大学发布的中文不确定性检测数据集语料上的实验,验证了本文提出的基于SVM的中文不确性信息检测方法的有效性,相比于句子评分模型,我们的系统取得了更好的召回率.

Automatic Identification of Uncertainty Information in Natural Language

There are a lot of uncertainty information in natural Language. It is becoming a focus of researches in NLP recently. However, the research on Chinese is still scarce. In this paper, we use SVM, which is a good solution to the high dimension, nonlinear and local small sample, to identify Chinese uncertainty information recognition as a question of classification. We carry experiment on Chinese uncertainty corpus published by Fudan University, which confirms the availability exposed by our paper based on SVM. Compared to the sentence scoring model, our system has better recall rate.