基于条件随机域模型的英语时间表达式识别研究  

English Temporal Expression Recognition Based on Conditional Random Fields

作　　者：王凤玲[1] 

机构地区：[1]菏泽学院计算机与信息工程系

出　　处：《电子技术（上海）》2012年第5期8-10,共3页Electronic Technology

摘　　要：时间表达式在多个自然语言处理领域都有着重要的作用,为了更好地识别时间表达式,提出使用条件随机域模型结合多种特征的方法对英语时间表达式进行识别,并采用TimeML标记语言对识别结果进行标记。采用Timebank1.1作为评测语料,通过实验结果发现各种特征的选择和应用是系统非常重要的部分,文中所选取的特征对于英语时间表达式的识别来说是非常成功的。Temporal expressions play very important role in multiple Natural Language Processing fields. In order to recognize Timexes more effectively, this paper adopts conditional random fields combining with many features to recognize the English temporal expression, and adopts the TimeML as the markup language to mark the recognition results. In the experiment sections, the paper uses Timebankl. 1 as the evaluation corpus. Experimental results show that the choice and application of various features are the key components for the system, and its result is crucial to the system performance. The characteristics chosen in the paper are very successful for recognizing the English temporal expressions.

关 键 词：时间表达式 条件随机域 TimeML 时间表达式识别与归范化(TERN) 特征 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]