A Correcting Model Based on Tribayes for Real-word Errors in English Essays

摘要：

This paper addresses the problem of real-word spelling errors, and also the problem of omission of effective features due to deficiency of training set in spelling correction. Then a method called RCW (real-word correction with WordNet) based on Tribayes is introduced, and it solves these problems to a certain extent. Drawing upon the context information, the score of ambiguous words are calculated and regarded as decisive factor for real-word errors correction in RCW. Moreover, the synonyms of the effective features ignored are extracted from WordNet, and we use them as feature so as to improve the accuracy of real-word errors correction. Experiment shows that RCW is able to provide a better performance than Microsoft Word 2007 on real-word errors correction.

关键词： Real-word errors Tribayes Omission of effective features WordNet Synonym

作者: Ya ZHOU Shenghao JING Guimin HUANG Shaozhong LIU Yan ZHANG

作者单位: Research Center on Data Science and Social Computing Guilin University of Electronic Technology Guilin 541004, China

会议类型: 国际会议

会议名称: 2012 Fifth International Symposium on Computational Intelligence and Design 第五届计算智能与设计国际会议 ISCID 2012

会议地点: 杭州

会议语种:英文

页码: 407-410

在线出版日期: 2012-10-28（万方平台首次上网日期，不代表论文的发表时间）