首页> 外文会议>International Conference on Machine Learning, Big Data, Cloud and Parallel Computing >Text Summarization using Partial Textual Entailment based Graphs
【24h】

Text Summarization using Partial Textual Entailment based Graphs

机译:基于部分文本意外的图形的文本摘要

获取原文

摘要

Information explosion has boosted research communities to propose many text summarization methods using different approaches. Text connectedness is a potential textual attribute to identify significant sentences from the source document to form a meaningful summary. A recent work that employs a graph based method where nodes represent sentences and where Textual Entailment (TE) is used to measure connectedness relationship between sentences. TE is a recent and good indicator of text inference in NLP applications. The minimum vertex cover method is used to generate cover of vertices that correlates to summaries of the document. The method has yielded good summaries. However, we have observed the quality of summarization is limited by the accuracy of TE determination. This paper attempts to overcome this limitation by employing Partial Textual Entailment (PTE) in the same. Therefore, this paper proposed a modified method which has shown competitive results on standard dataset.
机译:信息爆炸已经提高了研究社区,以提出使用不同方法的许多文本摘要方法。文本连接是一个潜在的文本属性,以识别来自源文档的重要句子,以形成有意义的摘要。最近的工作采用基于图形的方法,其中节点代表句子以及文本entailment(te)用于测量句子之间的连接关系。 TE是NLP应用中的初始和良好的文本推理指标。最小顶点覆盖方法用于生成与文档的摘要相关的顶点覆盖。该方法产生了良好的摘要。但是,我们已经观察到摘要质量受到TE测定的准确性的限制。本文试图通过使用相同的部分文本征报(PTE)来克服这种限制。因此,本文提出了一种修改方法,它在标准数据集上显示了竞争结果。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号