In this paper, a new powerful method in artificial neural networks, called modular network SOM (mnSOM) is introduced. mnSOM is a generalization of Self Organizing Maps (SOM) formed by replacing each vector unit of SOM with function module. The modular function could be a multi layer perceptron, a recurrent neural network or even SOM itself. Having this flexibility, mnSOM becomes a new powerful ...

| We show that the Fourier transform of the linear output of a single hidden layer perceptron consists of a multitude of line masses passing through the origin. Each line corresponds to one of the hidden neurons and its slope is determined by that neuron's weight vector. We also show that convolving the output of the network with a function can be achieved simply by modifying the shape of the s...

In recent years, many attempts have been made to predict the behavior of bonds, currencies, stocks, or stock markets. In this paper, the StandardlkPoors 500 Index is modeled using different neural network classification architectures. Most previous experiments used multilayer perceptrons for stock market forecasting. In this paper, a multilayer perceptron architecture and ZL probabilistic neura...

We propose a two-stage training for the multilayer perceptron (MLP). The first stage is bottom-up, where we use a class separability measure to conduct hidden layer training and the least squared error criterion to train the output layer. The second stage is top-down, we use a criterion derived from classification error rate to further train the network weights. We demonstrate the effectiveness...

The water quality of the Karaj River was studied through collecting 2137 experimental data set gained by 20 sampling stations. The data included different parameters such as T (temperature), pH, NTU (turbidity), hardness, TDS (total dissolved solids), EC (electrical conductivity) and basic anion, cation concentrations. In this study a multi-layer perceptron artificial neural network model was d...

Bio-absorbent palm fiber was applied for removal of cationic violet methyl dye from water solution. For this purpose, a solid phase extraction method combined with the artificial neural network (ANN) was used for preconcentration and determination of removal level of violet methyl dye. This method is influenced by factors such as pH, the contact time, the rotation speed, and the adsorbent dosag...

In software engineering there are plenty of applications used for reduced complexity and improved fault prediction approaches. In this paper we study various metrics that are not very much suitable to find fault classes in software. Basically using the concept of metrics to find fault classes and reduced complexity of classes. . various techniques like linear regression, logistic regression, on...

The loading problem is the problem to decide if a neural architecture can map a training set correctly with an appropriate choice of the weights. The following results will be shown: The loading problem is NP-complete for any feedforward perceptron architecture with at least two neurons in the rst hidden layer and varying input dimension. Further, it is NP-complete if the input dimension is xed...