The Research on Chinese Coreference Resolution Based on Maximum Entropy Model and Rules

  • Authors:

  • Yihao Zhang;Jianyi Guo;Zhengtao Yu;Zhikun Zhang;Xianming Yao

  • Affiliations:

  • The School of Information Engineering and Automation, Kunming University of Science and Technology, Kunming 650051;The School of Information Engineering and Automation, Kunming University of Science and Technology, Kunming 650051 and The Institute of Intelligent Information Processing, Computer Technology Appl ...;The School of Information Engineering and Automation, Kunming University of Science and Technology, Kunming 650051 and The Institute of Intelligent Information Processing, Computer Technology Appl ...;The School of Information Engineering and Automation, Kunming University of Science and Technology, Kunming 650051;The School of Information Engineering and Automation, Kunming University of Science and Technology, Kunming 650051

  • Venue:
  • WISM '09 Proceedings of the International Conference on Web Information Systems and Mining
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Coreference resolution is an important research topic in natural language processing, including the coreference resolution of proper nouns, common nouns and pronouns. In this paper, a coreference resolution algorithm of the Chinese noun phrase and the pronoun is proposed that based on maximum entropy model and rules. The use of maximum entropy model can integrate effectively a variety of separate features, on this basis to use rules method to improve the recall rate of digestion, and then use filtering rules to remove "noise" to further improve the accuracy rate of digestion. Experiments show that the F value of the algorithm in a closed test and an open test can reach 85.2% and 76.2% respectively, which improve about 12.9 percentage points and 7.8 percentage points compare with the method of rules respectively.