This paper presents a recognition system for English Handwritten that utilized Freeman Chain Code (FCC) as data representation. There are 544 features were extracted from character images that used six techniques to extract the features. Before extracting the features, thinning algorithm was applied to the original image to produce a Thinned Binary Image (TBI). A feed forward back propagation n...

This paper illustrates a novel method to analyze artificial neural networks so as to gain insight into their internal functionality. To this purpose, we will show analysis results of some feed-forward– error-back-propagation neural networks for image processing. We will describe them in terms of domain-dependent basic functions, which are, in the case of the digital image processing domain, dif...

In this paper, the effect of different alloying elements on the ultimate tensile strength of Al-Mg2Si composites is theoretically studied. The feed forward back propagation neural network with sigmoid function is used. The extensive experimental results taken from literature are modeled and mathematical formula is presented in explicit form. In addition, it is observed that magnesium and copper...

Recent theoretical advances in the learning of deep artificial neural networks have made it possible to overcome a vanishing gradient problem. This limitation has been overcome using a pre-training step, where deep belief networks formed by the stacked Restricted Boltzmann Machines perform unsupervised learning. Once a pre-training step is done, network weights are fine-tuned using regular erro...

The elements of the Hessian matrix consist of the second derivatives of the error measure with respect to the weights and thresholds in the network. They are needed in Bayesian estimation of network regularization parameters, for estimation of error bars on the network outputs, for network pruning algorithms, and for fast re-training of the network following a small change in the training data....

We present here a new model and algorithm which performs an efficient Natural gradient descent for Multilayer Perceptrons. Natural gradient descent was originally proposed from a point of view of information geometry, and it performs the steepest descent updates on manifolds in a Riemannian space. In particular, we extend an approach taken by the “Whitened neural networks” model. We make the wh...

In the present study, the abilities of three classification methods of data mining namely artificial neural networks with feed-forward back propagation algorithm, J48 decision tree method and logistic regression analysis are compared in a medical real dataset. The prediction of malignancy in suspected thyroid tumour patients is the objective of the study. The accuracy of the correct predictions...

This paper addresses the application of Self-adaptive Global Best Harmony Search (SGHS) algorithm for the supervised training of feed-forward neural networks (NNs). A structure suitable to data representation of NNs is adapted to SGHS algorithm. The technique is empirically tested and verified by training NNs on two classification benchmarking problems. Overall training time, sum of squared err...

Transmission and distribution lines are vital links between generating units and consumers. They are exposed to atmosphere, hence chances of occurrence of fault in transmission line is very high, which has to be immediately taken care of in order to minimize damage caused by it. In this paper discrete wavelet transform of voltage signals at the two ends of the transmission lines have been analy...

In this paper we show different evolutionary algorithms in order to optimise on-line weights of feed-forward neural networks when applied to short term (20 min.) urban traffic prediction. We compare the evolutionary methods with the classical back-propagation algorithm and we show results when weights are off-line and on-line evolved. Preliminary results are very promising and show the effectiv...