Pattern Acquisition for Chinese Named Entity Recognition: A Supervised Learning Approach

摘要

This paper presents a supervised learning method for the pattern acquisition for handcrafted rule-based Chinese named entity recognition systems. We automatically extracted low frequency patterns based on the predefined high-frequency patterns and manually validated the new patterns and outputs of terms. The experiments show that the number of person names extracted from the Chinese Treebank increased by 14.3% after the use of the new patterns.