一种基于词矢量的汉语语义量化模型

摘要

A word-vector-based Chinese word sense quantization model is proposed, which can be used to solve problems such as auto acquisition and quantization of word sense information. The modeling method of the model and its applications in the Chinese word sense disambiguation are further described. And then, the model's ability to discriminate word sense is evaluated by constructing pseudoword. The experiment shows that this model has a good representation for word sense. As the construction of model is done via the statistic of large-scale rough corpora, huge workload of manually quantization word senses is avoided. So this model can be applied to many word-sense-related NLP tasks.%通过建立基于词矢量的汉语语义量化模型来解决语义信息的自动获取及量化问题，描述了模型的建立方法及其在汉语词义排歧中的应用，最后通过构造伪词的方法对模型的语义辨识能力进行了评测.实验表明该语义量化模型具有很好的语义表示能力，并且由于模型的建立是通过对大规模生语料库的统计来完成的，避免了人工对词语语义进行量化时所需的庞大工作量，从而可以运用于许多与语义相关的自然语言处理任务中.