Wavelet Probabilistic Neural Network Research Articles

Wavelet Probabilistic Neural Networks.

In this article, a novel wavelet probabilistic neural network (WPNN), which is a generative-learning wavelet neural network that relies on the wavelet-based estimation of class probability densities, is proposed. In this new neural network approach, the number of basis functions employed is independent of the number of data inputs, and in that sense, it overcomes the well-known drawback of traditional probabilistic neural networks (PNNs). Since the parameters of the proposed network are updated at a low and constant computational cost, it is particularly aimed at data stream classification and anomaly detection in off-line settings and online environments where the length of data is assumed to be unconstrained. Both synthetic and real-world datasets are used to assess the proposed WPNN. Significant performance enhancements are attained compared to state-of-the-art algorithms.

Artificial neural network approach for locating internal faults in salient-pole synchronous generator

This paper presents an artificial neural network approach for locating internal faults in salient-pole synchronous generator. This method uses samples of magnetic flux linkages to reach a decision. Meyer wavelet probabilistic neural network (MWPNN) as main part of this fault diagnosis method is utilized to detect internal faults. The basic PNN is integrated with discrete wavelet transform (DWT) to construct the MWPNN. The MWPNN is trained by features extracted from the magnetic flux linkage data through the discrete wavelet transform (DWT). To implement this method practically, the training database for the PNN is generated by 3D-FEM. Since the FEM simulation is a time-consuming process, hence in order to reduce the number of FEM computations generalized regression neural network (GRNN) is used to estimate the flux linkages characteristics of synchronous generator under various conditions. The diagnosis performance of the proposed method is validated via experimental data, derived from a 4-pole, 380V, 1500rpm, 50Hz, 50KVA, 3-phase salient-pole synchronous generator.

Hindi paired word recognition using probabilistic neural network

Automatic speech recognition has been a subject of active research interest since last few decades. In the present paper, spoken Hindi (Indian national language) Paired Word Recognition (HPWR) has been examined with the help of intelligent hybrid computing scheme based on wavelet transform and Probabilistic Neural Network (PNN). This type of network is a combination of radial basis layer and a competitive transfer function layer, which picks up the maximum probabilities as a final result. For the experimental purpose, six hundred and fifty Hindi paired word samples from individuals with different gender and age groups have been recorded. Pre-processing procedure has been performed on the samples using accurate endpoint detection algorithm. For feature extraction of samples, wavelet transform has been used. PNN algorithm is used as a classifier. The proposed intelligent hybrid computing scheme based on wavelet-probabilistic neural network has produced practically good recognition rate.