论文部分内容阅读
根据互信息、RBF神经网络和关联规则原理,提出一种抽取Web文本分类规则的新方法。首先根据互信息选择相关程度大的若干词条,然后采用RBF神经网络方法对选择的特征作进一步提取,得到维数较小的文本特征向量空间,最后根据挖掘出的关联规则获取Web文本分类规则,建立文本分类器,在保证分类精度的前提下,抽取出利于理解的文本分类规则。
According to the principle of mutual information, RBF neural network and association rules, a new method of extracting Web text classification rules is proposed. First of all, according to the mutual information, we select several terms with relative relevance, and then use the RBF neural network method to further extract the selected features to get the dimension of the text eigenvector space. At last, the Web text classification rules are obtained according to the mining association rules , The establishment of text classifier, in ensuring the accuracy of classification under the premise of extracting text classification rules conducive to understanding.