Computer and Modernization

Previous Articles    

Chinese Agricultural Named Entity Recognition Based on Conditional Random Fields

  

  1. College of Information Engineering, Yangzhou University, Yangzhou 225000, China
  • Received:2017-05-04 Online:2018-01-23 Published:2018-01-24

Abstract: To meet the requirement of diversification, the paper introduces named entities model based on conditional random fields, divides agricultural named entities into 4 classes, which are diseases and pests, crops, chemical fertilizer, pesticide and annotates. These entities are labeled with custom annotation set. It is segmented collected corpus with ICTCLAS segment system and added features to raise the recognition rate. Ultimately, the accuracy rate reaches over 93% and the recalling rate reaches over 84%. This result confirms that recognizing multiple entities after classifying them is feasible.

Key words: Chinese named entity recognition, conditional random field, agriculture

CLC Number: