There are some deficiencies in the improved three-ratio method even though it has been widely used in power transformer fault diagnosis. Using artificial neural networks, the power transformer fault diagnosis is improved in this article. With Matlab programming, three different kinds of neural networks, which are Radial Basis Function (RBF) neural network, Learning Vector Quantization (LVQ) neu...

This paper compares three mathematical models for surgical decisions on head injury patients. A logistic regression and two neural network models were developed using a large clinical database. Using randomly selected 9480 cases as the training group and another 3160 cases as the validation group. We evaluated the performance of a logistic regression model, a multi-layer perceptron (MLP) neural...

One of the main obstacles to the widespread use of artificial neural networks is the difficulty of adequately defining values for their free parameters. This work discusses how Radial Basis Function (RBF) neural networks can have their free parameters defined by Genetic Algorithms (GAs). For such, it firstly presents an overall view of the problems involved and the different approaches used to ...

Neural network process modelling needs the use of experimental design and studies. A new neural network constructive algorithm is proposed. Moreover, the paper deals with the influence of the parameters of radial basis function neural networks and multilayer perceptrons network in process modelling. Particularly, it is shown that the neural modelling, depending on learning approach, cannot be a...

In this paper it is investigated whether neural networks are able to improve the performance of a PI controller when controlling a combustion engine. The idea is not to replace but to assist a PI controller by a neural co-controller. Three different neural approaches are investigated for this use: Dynamic RBF (DRBF), Adaptive Time-Delay Neural Network (ATNN), and Local Ellipsoidal Model Network...

Yaobin Qin [email protected] Supervisor: Pro.lilja Department of Electrical and Computer Engineering Abstract Neural networks are family statistical learning algorithms and structures and are used to estimate or approximate functions and pattern classification. The Neural network system is constructed through interconnected neurons and training weights. The paper will present the improvement of ...

A complex radial basis function neural network is proposed for equalization of quadrature amplitude modulation (QAM) signals in communication channels. The network utilizes a sequential learning algorithm referred to as complex minimal resource allocation network (CMRAN) and is an extension of the MRAN algorithm originally developed for online learning in real-valued radial basis function (RBF)...

Scatterograms of images of training vectors in the hidden space help to evaluate the quality of neural network mappings and understand internal representations created by the hidden layers. Visualization of these representations leads to interesting conclusions about optimal architectures and training of such networks. The usefulness of visualization techniques is illustrated on parity problems...

We propose a genetic algorithm for generating adversarial examples for machine learning models. Such approach is able to find adversarial examples without the access to model’s parameters. Different models are tested, including both deep and shallow neural networks architectures. We show that RBF networks and SVMs with Gaussian kernels tend to be rather robust and not prone to misclassification...