作者:
G Mirchandani,W Cao
关键词:
general and miscellaneous//mathematics; computing; and inf...;parallel processing;computer networks;array processors;computer calculations;learning;nerve cells;simulation;training;animal cells
摘要:
Recent results indicate that the number of hidden nodes (H) is a feedforward neural net depend only on the number of input training patterns (T). There appear to be conjectures that H is of the order of (T-1) and of log/sub 2/T. The authors present proof that maximum number of separable regions (M) in the input space is a function of both H and input space dimension (d). They also show that H = m - 1 and H = log/sub 2/M are special cases of that formulation. M defines a lower bound on T, the number of input patterns that may be used for training. Application to some experiments are investigated.
在线下载