Constructive proof of efficient pattern storage in the multi-layer perceptron

Abstract

In this paper, we

Forecast the Dangerous Level of Gas Outburst Based on GASA Neural Networks in Coal Mine

“…In OWO-BP which uses OWO to update output weights and BP to update hidden weights, linear equations are solved to find output weights and BP is used to find hidden weights [13]. In our approach we have adapted the idea [14] of minimizing a separate error function for each hidden unit to find the hidden weights, termed as HWO.…”

“…It is a complex and nonlinear mapping or identification problem to forecast coal and gas outburst correctly. Various prediction techniques such as experience and knowledge-based method and neural nets (NN) [3], [5], [9], [10], [13] have been proposed. Nonlinear system identification via neural networks consists in adjusting the network in such a way that it approximately describes the inputoutput mapping of the system.…”

Review article: Attributes of neural networks for extracting continuous vegetation variables from optical and radar measurements

“…( 2) once trained, the MLP can be applied to data one or more orders of magnitude faster than the NNE's; ( 3) a MLP can closely approximate the performance of a NNE if it can process the NNE's cluster vectors without error (memorize the clusters and their associated output vectors); and ( 4 ) a MLP or Volterra ® lter can memorize as many patterns as it has free parameters per output node (its complexity) (Gopalakrishnan et al 1994).…”

A neural network training algorithm utilizing multiple sets of linear equations

“…Unfortunately, backpropagation is not a very effective method for updating hidden weights [15,29]. Some researchers [11,16,17,20,31] have used the Levenberg-Marquardt(LM) method to train the multilayer perceptron.…”