Comparison of Two Methods Used to Data Preprocessing for One Kind of Neura l Network
Yao Hong · Microcomputer Development · 2003
Presents the basic algorithms of how to use principal compo nent analysis (PCA )and rough sets (RS) to compress original datas.A group o f slope engineer data,after preprocessing with PCA and RS respectively, are giv en to a BP neural network so that the slope state output is approximated. The si mulation of test samples shows that RS have advantages over PCA in this situatio n.