论文部分内容阅读
【目的/意义】文本自动摘要能快速获取文本主要内容,极大提高信息使用效率,帮助人们从信息海洋中解放出来。随着互联网大数据日益深入发展,文本信息的数量已经远远超出人工处理极限,文本自动摘要研究显得越发迫切和重要。【方法/内容】本文通过对过去70年国内外文本自动摘要经典文献重要文献进行收集、整理和分析,总结归纳出六类主要文本自动摘要方法及其理念和具体做法,对比评析其优势不足,并对未来研究方向进行展望,绘制出该研究领域的一个发展全景图。【结果/结论】自动摘要方法所使用的特征经历了由简单到复杂、由个体到联系、由表层到深层的发展路径;如何深入利用深度学习方法,以及如何将传统方法与深度学习方法更好结合起来将是下一步研究的热点和重点。
[Purpose / Significance] Automatic Text Digest Fast access to the main text of the text, greatly improving the efficiency of the use of information to help people liberated from the ocean of information. With the development of Internet big data, the quantity of textual information has been far beyond the manual processing limit. It is more and more urgent and important to research the text automatic digest. 【Method / Content】 This essay collects, collate and analyze the important literature of the classic literature of automatic summarization of texts at home and abroad in the past 70 years, summarizes the six main methods of automatic summarization of texts and their concepts and specific practices, compares and analyzes their advantages, And look forward to the future research direction, draw a development panorama of the research area. [Results / Conclusions] The features used in the automatic summarization approach have undergone a path from simple to complex, from individual to contact, from surface to deep; how to use deep learning methods in depth and how to make traditional methods and depth learning methods better Combined with the next step will be the focus of research.