PROJECT TITLE :

Bidirectional Extreme Learning Machine for Regression Problem and Its Learning Effectiveness

ABSTRACT:

It's clear that the educational effectiveness and learning speed of Neural Networks are in general so much slower than needed, that has been a serious bottleneck for several applications. Recently, a simple and efficient learning method, referred to as extreme learning machine (ELM), was proposed by Huang , which has shown that, compared to some standard methods, the training time of neural networks can be reduced by a thousand times. But, one in every of the open issues in ELM research is whether the quantity of hidden nodes can be any reduced without affecting learning effectiveness. This transient proposes a replacement learning algorithm, referred to as bidirectional extreme learning machine (B-ELM), in that some hidden nodes aren't randomly selected. In theory, this algorithm tends to cut back network output error to 0 at an extraordinarily early learning stage. Furthermore, we find a relationship between the network output error and also the network output weights in the proposed B-ELM. Simulation results demonstrate that the proposed technique will be tens to many times faster than different incremental ELM algorithms.