Abstract: In order to reduce the influences of the uncertainty in natural language to the traditional text classification method, this paper puts forward a new text classification method combining with cloud model, which defines cloud model of document and category, through computing the cloud similarity between test document and each category. The test text is assigned to a category based on maximum similarity principle. To sum up, experimental results show that the classification accuracy has improved by using this method, compared with the K-NN.

Key words: text classification, cloud model, cloud similarity

摘要： 为了降低在传统的文本分类方法中自然语言的不确定性对分类效果的影响，提出了一种结合云模型的文本分类方法。该方法分别定义文本和类别的云模型，通过计算测试文本和每个类别的云相似度，根据最大相似度原则确定测试文本所属的类别。实验结果表明，与传统的K-NN算法相比，该方法在分类准确率等方面有所提高。

关键词: 文本分类, 云模型, 云相似度