Algorithm

        YELM, a modified extreme learning machine, is an algorithm, using Sigmoidal function as activation function, which trains a single-hidden layer forward neural network (SFNN).

Matlab codes

        YELM Demo for Diabetes Version 1.0

Algorithm features

• YELM trains a single hidden layer forward neural network with Sigmoidal function as activation function

• Using rank-1 and rank-2 methods to obtain the optimal estimate for the output weights when column and row ranks of the hidden layer output matrix are not full

References

Y. Yuan, Y. G. Wang, F. Cao. Optimization approximation solution for regression problem based on extreme learning machine. Neurocomputing, 74(16):2475--2482, 2011.

« Back