Statistical Methods to Evaluate Important Degrees of Document Features

摘要：

　　Documents are often analyzed based on features.Evaluation of important degrees of features is a basic work of document analysis.Now this work is done mainly by human,so the workload is too large and results are subjective.Three statistical methods are proposed based on the actual usage of features in a large number of documents.The statistical values and those from experts are contrasted and analyzed,and then a selection result based on voting is given.Some results from 500 document samples are given.

关键词： document format document feature statistics document processing

作者: Xia HOU Ning LI Hong-bo YANG

作者单位: Computer School,Beijing Information Science and Technology University,Beijing,China;Beijing Key Labo Computer School,Beijing Information Science and Technology University,Beijing,China Experimental Teaching Center of Electronics information and Control,Beijing Information Science and

会议类型: 国内会议

会议名称: 2014全国文档信息处理学术会议

会议地点: 北京

会议语种:英文

页码: 1-10

在线出版日期: 2014-11-01（万方平台首次上网日期，不代表论文的发表时间）