基于语义网络的中文词汇链构造方法

Constructing Chinese Lexical Chains Based on Semantic Network

词汇链是文本中一系列词汇关联而成的语义链,构造词汇链有助于提高文本的连贯性,把握文本主题。借鉴概念层次网络(HNC)理论对词汇语义进行形式化描述的原理,研究基于语义网络节点遍历计算的词汇链构造方法。首先对汉语文本进行分词处理,从知识库中获得词语的HNC符号,然后对语义符号进行解析,获得词语的语义关联性并构造词汇链;最后,对词汇链进行优选处理,设定阈值获得优选词汇链。实验结果表明,该方法构造的词汇链可接受度较高,对主题标引、自动文摘、文本分类等有重要价值。

Lexical chain is a series of words associated with the semantic link in the text, by which improves the coherence and grasps the text topic. Based on Hierarchical Network of Concepts（ HNC ） formal description of the lexical semantics, lexical chains construction will be converted to a semantic network nodes traversal. Firstly, the word segmentation of Chinese text will be processed in advance, and the HNC symbols will be assigned to each word. Then, semantic relevance behind the HNC symbols will be obtained to construct lexical chains after parsing. Finally, lexical chains can be optimized according to the threshold. Experimental results show that the lexical chains can be constructed with higher acceptability, which is valuable to subject indexing, automatic summarization and text classification.