scholarly journals Estimating Weibull Parameters Using Least Squares and Multilayer Perceptron vs. Bayes Estimation

2022 ◽  
Vol 71 (2) ◽  
pp. 4033-4050
Author(s):  
Walid Aydi ◽  
Fuad S. Alduais
Author(s):  
A. Zabidi ◽  
W. Mansor ◽  
Khuan Y. Lee

<p>Mel Frequency Cepstral Coefficient is an efficient feature representation method for extracting human-audible audio signals. However, its representation of features is large and redundant. Therefore, feature selection is required to select the optimal subset of Mel Frequency Cepstral Coefficient features. The performance of two types of feature selection techniques; Orthogonal Least Squares and F-ratio for selecting Mel Frequency Cepstral Coefficient features of infant cry with asphyxia was examined. OLS selects the feature subset based on their contribution to the reduction of error, while F-Ratio selects them according to their discriminative abilities. The feature selection techniques were combined with Multilayer Perceptron to distinguish between asphyxiated infant cry and normal cry signals. The performance of the feature selection methods was examined by analysing the Multilayer Perceptron classification accuracy resulted from the combination of the feature selection techniques and Multilayer Perceptron. The results indicate that Orthogonal Least Squares is the most suitable feature selection method in classifying infant cry with asphyxia since it produces the highest classification accuracy.<em></em></p>


1995 ◽  
Vol 7 (5) ◽  
pp. 982-999 ◽  
Author(s):  
Mikko Lehtokangas ◽  
Jukka Saarinen ◽  
Kimmo Kaski ◽  
Pentti Huuhtanen

Usually the training of a multilayer perceptron network starts by initializing the network weights with small random values, and then the weight adjustment is carried out by using an iterative gradient descent-based optimization routine called backpropagation training. If the random initial weights happen to be far from a good solution or they are near a poor local optimum, the training will take a lot of time since many iteration steps are required. Furthermore, it is very possible that the network will not converge to an adequate solution at all. On the other hand, if the initial weights are close to a good solution the training will be much faster and the possibility of obtaining adequate convergence increases. In this paper a new method for initializing the weights is presented. The method is based on the orthogonal least squares algorithm. The simulation results obtained with the proposed initialization method show a considerable improvement in training compared to the randomly initialized networks. In light of practical experiments, the proposed method has proven to be fast and useful for initializing the network weights.


Sign in / Sign up

Export Citation Format

Share Document