An approach to feature dimensionality reduction based on radial basis function neural networks

作者：Li, Tao^*; Xiao, Nanfeng^*

来源：Journal of Convergence Information Technology, 2012, 7(3): 117-126.

DOI：10.4156/jcit.vol7.issue3.14

• Data classification
• Decision tree classification
• Dimensionality reduction
• Dimensionality reduction method
• Features
• High dimensional data
• Priori knowledge
• Radial basis function neural networks

In this paper we introduce a new feature dimensionality reduction method for decision tree classification. We combine the neural network with decision tree, first, the paper analyzes the problem about dimensionality reduction, sorts the feature value in importance order in the first place, second, we take advantage of the black box classification characteristic of neural network which does not need a priori knowledge, and its high-efficiency in classification, prune high dimensional data one by one, select several most effective basic features for data classification, we get the effect of dimensionality reduction by using this method.

• 出版日期2012-2
• 单位华南理工大学; 华南农业大学