Skip to main navigation Skip to search Skip to main content

Rejection algorithm for mixed Chinese/English outliers

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

This paper introduces several algorithms of character outlier rejection, which include algorithm based on estimation of recognition confidence and algorithm based on ONE-CLASS SVM. As for the former, confidence estimation of classifier based on distance is discussed, which includes algorithms based on Bayes and empirical rules. Lots of experiments are done based on different rejection algorithms, and the results are given in the end. Some key problems are presented and some solutions are recommended.

Original languageEnglish
Title of host publicationProceedings - 2010 IEEE International Conference on Intelligent Computing and Intelligent Systems, ICIS 2010
Pages415-418
Number of pages4
DOIs
StatePublished - 2010
Externally publishedYes
Event2010 IEEE International Conference on Intelligent Computing and Intelligent Systems, ICIS 2010 - Xiamen, China
Duration: 29 Oct 201031 Oct 2010

Publication series

NameProceedings - 2010 IEEE International Conference on Intelligent Computing and Intelligent Systems, ICIS 2010
Volume3

Conference

Conference2010 IEEE International Conference on Intelligent Computing and Intelligent Systems, ICIS 2010
Country/TerritoryChina
CityXiamen
Period29/10/1031/10/10

Keywords

  • Bayes posteriori probability
  • Character outlier rejection
  • Empirical rules
  • ONE-CLASS SVM

Fingerprint

Dive into the research topics of 'Rejection algorithm for mixed Chinese/English outliers'. Together they form a unique fingerprint.

Cite this