摘要
概率潜在语义模型使用统计的方法描述"文档—潜在语义—词"之间的概率分布关系,其实质是模拟了潜在的概率语义空间,并将文档和词映射到同一个语义空间.该文将概率潜在语义分析模型用于中文文本分类,一方面较好地处理了自然语言中的同义、多义问题;另一方面通过计算概率潜在语义空间中向量的距离来获得文档间的类别信息从而达到文本分类的目的.实验结果表明,该分类器具有良好的分类性能.
The model of probability latent semantic analysis based on statistical methods describes the probability distribution between latent semantics and documents or words,by which a probabilistic latent semantic space is actually simulated.The words and documents are mapped into the same semantic space.In this paper,the model is applied to Chinese text classification.On one hand,it deals with the synonyms and polysemy in natural language better.On the other hand,the class information among documents is obtained by calculating the vector distance in probabilistic latent semantic space,so as to achieve the purpose of text classification.The results show that the classifier has a good performance.
出处
《甘肃联合大学学报（自然科学版）》
2011年第4期75-78,共4页
Journal of Gansu Lianhe University :Natural Sciences
关键词
中文文本分类
概率潜在语义分析
语义空间
Chinese text classification
probabilistic latent semantic analysis（PLSA）
semantic space
