首页 | 本学科首页   官方微博 | 高级检索  
     

文本分类器性能评估指标
引用本文:宋枫溪,高林.文本分类器性能评估指标[J].计算机工程,2004,30(13):107-109,127.
作者姓名:宋枫溪  高林
作者单位:1. 炮兵学院1系,合肥,230031;南京理工大学计算机系,南京,210094
2. 南京理工大学计算机系,南京,210094
摘    要:讨论了用传统性能评估指标(精度和错误率)评估文本分类器时面临的困难,分析了目前几种常用的文本分类器性能评估指标:查全率-查准率曲线、AUROC、Fi值以及BEP值的优点与不足,并提出了两种新的可用于文本分类器的性能评估指标。

关 键 词:文本分类  性能评估  BEP值  ROC曲线
文章编号:1000-3428(2004)13-0107-03

Performance Evaluation Metric for Text Classifiers
SONG Fengxi,GAO Lin.Performance Evaluation Metric for Text Classifiers[J].Computer Engineering,2004,30(13):107-109,127.
Authors:SONG Fengxi  GAO Lin
Affiliation:SONG Fengxi,GAO Lin 1,2 2
Abstract:Afterdiscussing on the deficiency of accuracy and error rate as performance metrics for text categorization problems, this paper analyzes the strength and shortcoming of recall-precision curve, area under the ROC curve, F1value and break-even point, which are the most prevailing performance metrics for text classifiers, and proposes two kinds of new performance metric for text classifier.
Keywords:Text categorization  Performance evaluation  Breaks-even point  Receiver operating characteristic curve
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号