Part of Neural Information Processing Systems 0 (NIPS 1987)
G.Z. SUN, Y.C. LEE, H.H. CHEN
We propose a new scheme to construct neural networks to classify pat(cid:173)
terns. The new scheme has several novel features :
We focus attention on the important attributes of patterns in ranking order. Extract the most important ones first and the less important ones later.
In training we use the information as a measure instead of the error