Automated error detection and correction of chinese characters in written essays based on weighted finite-state transducer

Shudong Hao; Zongtian Gao; Mingqing Zhang; Yanyan Xu; Hengli Peng; Kaile Su; Dengfeng Ke

doi:10.1109/ICDAR.2013.156

Automated error detection and correction of chinese characters in written essays based on weighted finite-state transducer

Shudong Hao
, Zongtian Gao
, Mingqing Zhang
, Yanyan Xu
, Hengli Peng
, Kaile Su
, Dengfeng Ke

Department of Computer Science

Beijing Forestry University
Beijing Language and Culture University
Zhejiang Normal University
Chinese Academy of Sciences

Research output: Contribution to journal › Conference article › peer-review

4 Scopus citations

Abstract

Chinese text error detection and correction is widely applicable, but the methods so far are not robust enough for industrial use. In this paper, a new method is proposed based on Tri-gram modeled-Weighted Finite-State Transducer (WFST). By integrating confusing-character table, beam search and A* search, we evaluate the performance on real test essays. Various experiments have been conducted to prove that the proposed method is effective with the recall rate of 85.68%, the detection accuracy of 91.22% and the correction accuracy of 87.30%.

Original language	English
Article number	6628721
Pages (from-to)	763-767
Number of pages	5
Journal	Proceedings of the International Conference on Document Analysis and Recognition, ICDAR
DOIs	https://doi.org/10.1109/ICDAR.2013.156
State	Published - 2013
Event	12th International Conference on Document Analysis and Recognition, ICDAR 2013 - Washington, DC, United States Duration: 25 Aug 2013 → 28 Aug 2013

Keywords

Error correction
Error detection
N-gram language model
Weighted Finite-State Transducer (WFST)

Access to Document

10.1109/ICDAR.2013.156

Cite this

@article{267e56da0f144cd180baeba4664baf47,

title = "Automated error detection and correction of chinese characters in written essays based on weighted finite-state transducer",

abstract = "Chinese text error detection and correction is widely applicable, but the methods so far are not robust enough for industrial use. In this paper, a new method is proposed based on Tri-gram modeled-Weighted Finite-State Transducer (WFST). By integrating confusing-character table, beam search and A* search, we evaluate the performance on real test essays. Various experiments have been conducted to prove that the proposed method is effective with the recall rate of 85.68\%, the detection accuracy of 91.22\% and the correction accuracy of 87.30\%.",

keywords = "Error correction, Error detection, N-gram language model, Weighted Finite-State Transducer (WFST)",

author = "Shudong Hao and Zongtian Gao and Mingqing Zhang and Yanyan Xu and Hengli Peng and Kaile Su and Dengfeng Ke",

year = "2013",

doi = "10.1109/ICDAR.2013.156",

language = "English",

pages = "763--767",

note = "12th International Conference on Document Analysis and Recognition, ICDAR 2013 ; Conference date: 25-08-2013 Through 28-08-2013",

}

TY - JOUR

T1 - Automated error detection and correction of chinese characters in written essays based on weighted finite-state transducer

AU - Hao, Shudong

AU - Gao, Zongtian

AU - Zhang, Mingqing

AU - Xu, Yanyan

AU - Peng, Hengli

AU - Su, Kaile

AU - Ke, Dengfeng

PY - 2013

Y1 - 2013

N2 - Chinese text error detection and correction is widely applicable, but the methods so far are not robust enough for industrial use. In this paper, a new method is proposed based on Tri-gram modeled-Weighted Finite-State Transducer (WFST). By integrating confusing-character table, beam search and A* search, we evaluate the performance on real test essays. Various experiments have been conducted to prove that the proposed method is effective with the recall rate of 85.68%, the detection accuracy of 91.22% and the correction accuracy of 87.30%.

AB - Chinese text error detection and correction is widely applicable, but the methods so far are not robust enough for industrial use. In this paper, a new method is proposed based on Tri-gram modeled-Weighted Finite-State Transducer (WFST). By integrating confusing-character table, beam search and A* search, we evaluate the performance on real test essays. Various experiments have been conducted to prove that the proposed method is effective with the recall rate of 85.68%, the detection accuracy of 91.22% and the correction accuracy of 87.30%.

KW - Error correction

KW - Error detection

KW - N-gram language model

KW - Weighted Finite-State Transducer (WFST)

UR - https://www.scopus.com/pages/publications/84889568099

UR - https://www.scopus.com/pages/publications/84889568099#tab=citedBy

U2 - 10.1109/ICDAR.2013.156

DO - 10.1109/ICDAR.2013.156

M3 - Conference article

AN - SCOPUS:84889568099

SN - 1520-5363

SP - 763

EP - 767

JO - Proceedings of the International Conference on Document Analysis and Recognition, ICDAR

JF - Proceedings of the International Conference on Document Analysis and Recognition, ICDAR

M1 - 6628721

T2 - 12th International Conference on Document Analysis and Recognition, ICDAR 2013

Y2 - 25 August 2013 through 28 August 2013

ER -

Automated error detection and correction of chinese characters in written essays based on weighted finite-state transducer

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this