基于统计语义和结构特征的自动文摘  

Automatic Summarization Based on Statistic Semantic and Structural Features

作　　者：黄丽琼[1] 何中市[1,2] 

机构地区：[1]重庆大学计算机学院 [2]重庆大学语言认知及信息处理研究所,重庆400044

出　　处：《广西师范大学学报（自然科学版）》2006年第4期187-190,共4页Journal of Guangxi Normal University:Natural Science Edition

基　　金：国家自然科学基金资助项目(60173060)

摘　　要：在分析自动文摘现有方法优缺点的基础上,提出了一种基于统计、语义和结构特征的自动文摘方法。用这些特征构成句子向量表示,并用机器学习的方法对其进行训练得到器,从而把自动文摘转换为分类问题。实验表明,该方法具有较好的重合率。同时,为了解决文摘的冗余和不连贯缺点,进行了一系列的后期处理,提高了文摘的质量。This paper presents a new automatic summarization method based on statistic, semantic and structural features while the advantages and disadvantages are analyzed for the popular methods of automatic summarization. There are eight features used to form the feature vector for each sentence,and the summarizer is gained by machine learning algorithms ,so automatic summarization is changed into classification task. The experiment results show that the method maintains higher precision. Meanwhile,the paper processes a series of post processing to overcome the shortcoming of redundancy and incoherence, and it improves largely the quality of summary.

关 键 词：机器学习 自动文摘 句子相似度 自然语言处理 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]