基于双向LSTM与CRF融合模型的否定聚焦点识别.PDFVIP

基于双向LSTM 与CRF 融合模型的否定聚焦点识别 沈龙骧，邹博伟，叶静，周国栋，朱巧明 （苏州大学计算机科学与技术学院，江苏 苏州 215006） 摘要：否定表达作为自然语言文本中常见的语言现象，对自然语言处理上层应用，如情感分析、 信息抽取等，具有十分重要的意义。否定聚焦点识别任务是更细粒度的否定语义分析，其旨在识别出 句子中被否定词修饰和强调的文本片段。本文将该任务作为序列标注问题，提出了一种基于双向长短 期记忆网络结合条件随机场（BiLSTM-CRF ）的否定聚焦点识别模型，其中，BiLSTM 网络能够充分利 用上下文信息并抓取全局特征，CRF 层能够有效学习输出标签之间的前后依赖关系。在*SEM2012 评 测任务数据集上的实验结果表明，基于BiLSTM-CRF 的否定聚焦点识别方法的准确率（Accuracy ）达 到69.58%，与目前最好的系统相比，性能提升了2.44% 。 关键词：否定聚焦点；BiLSTM-CRF 模型；序列标注 中图分类号：TP391 文献标识码：A Negation Focus Identification via Bi-directional LSTM-CRF Model SHEN Longxiang , ZOU Bowei , YE Jing , ZHOU Guodong and ZHU Qiaoming (School of Computer Science and Technology, Soochow University, Suzhou, Jiangsu 215006, China) Abstract : Negative expressions are common phenomena in natural language text and play a critical role in various applications of natural language processing, such as sentiment analysis, information extraction. Negation focus identification task is a finer-grained negative semantic analysis, which aims at identifying the text fragment modified and emphasized by a negative keyword. We regard the negation focus identification as a sequence labeling task, and propose a bidirectional Long Short-Term Memory network with a Conditional Random Field layer (BiLSTM -CRF). Such model can not only learn the contextual information bi-directionally and capture the global features by the BiLSTM network, but also learn the dependency between the output tags by the CRF layer. Experimental results on the *SEM2012 dataset shows that the performance of our approach achieves the accuracy of 69.58% with the improvement of 2.44%, compared to the state-of-the-art system for negation focus identification. Key words: Negation focus; BiLSTM-CRF model; Sequence labeling 1 引言 语义进行了反转。否定聚焦点是指在否定表