基于聚簇样本约减的K-近邻神经网络分类器  

K-Nearest Neighbor Neural Network Classifier of Samples Reduction Based on Clustering

作　　者：王春艳[1] 程霜梅[1] 杨鑫[1] 

机构地区：[1]长春师范学院计算机科学与技术学院,吉林长春130032

出　　处：《情报科学》2010年第10期1547-1549,1553,共4页Information Science

摘　　要：针对K-nn算法应用于文本分类时存在的问题进行改进,提出利用聚簇进行训练样本约减的方法,并用神经网络实现改进的k-nn分类器。实验表明,分类所用的时间比传统k-nn算法有明显减少,一定程度上提高了分类效率,同时分类准确性有所提高。A samples reduction based on clustering method for k-nn algorithms in text classification was presented to accelerate speed. The neural network is proposed in improved k-nn classifier. The experiment results show that the time of this method is less than traditional k-nn algorithms. At the same time, it improves the efficiency and precision of text classification to some degree.

关 键 词：文本分类 聚簇 k-nn算法 神经网络 

分 类 号：G350[文化科学—情报学]