汉语文本中特殊符号串的自动识别技术

Technique of Special Strings Automatic Recognition in Chinese Texts

提出从组成形式和上下文语境两个方面来自动识别汉语文本中的各种特殊符号串。其组成形式用包含约束式的上下文无关文法来描述,改进的LR分析方法进行形式识别;上下文语境采用基于知网概念的特征向量来表达,向量间的欧式距离表示语境间的相似度。实践证明该技术方案是相当有效的。

This paper puts forward two ways: the formalization and context to recognize the special strings in Chinese texts. Context free grammar extended by constrained formula and an improved LR parser are adopted to formalize and recognize the special strings. The context is represented by the eigenvector based on Hownet and their similarity is measured by Euclid distance between their eigenvectors. This technique is proved to be quite effective in practice.