Neural Networks and Fuzzy Systems

Abe, Shigeo

doi:10.1007/978-1-4615-6253-5

Cited by 37 publications

(14 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The learning process automatically adjusts the weights and thresholds of the processing elements in an effort to minimize the differences between the ANN output and the targeted output. This process is called training and is based on several different learning rules [7,8], structures, and types. Two of the simplest ones, most commonly used powerful and effective ANNs, are the multilayer perceptron (MLP) neural network and the radial basis function neural network (RBF) [7,8].…”

Section: Artificial Neural Network' (Anns) Implementationmentioning

confidence: 99%

“…This process is called training and is based on several different learning rules [7,8], structures, and types. Two of the simplest ones, most commonly used powerful and effective ANNs, are the multilayer perceptron (MLP) neural network and the radial basis function neural network (RBF) [7,8]. Both ANNs have been widely used, the last years, in the solution of many power system problems presenting very accurate results.…”

Section: Artificial Neural Network' (Anns) Implementationmentioning

confidence: 99%

See 1 more Smart Citation

An artificial neural network software tool for the assessment of the electric field around metal oxide surge arresters

Ekonomou

Mladenov

2015

Neural Comput & Applic

View full text Add to dashboard Cite

The paper presents an artificial neural network (ANN) software tool that has been developed in order to assess the electric field around medium-voltage surge arresters. The knowledge of the electric field around gapless metal oxide surge arresters is very useful for diagnostic tests and design procedures. For the training, validating, and testing of the ANNs, real data have been used collected from hundreds of measurements. The developed ANN software can be used by electric utilities in order to diagnose the condition of the surge arresters without stopping their operation and by laboratories and manufacturing/retail companies dealing with medium-voltage surge arresters and either face a lack of suitable measuring equipment or want to compare/verify their own measurements.

show abstract

Section: Artificial Neural Network' (Anns) Implementationmentioning

confidence: 99%

Section: Artificial Neural Network' (Anns) Implementationmentioning

confidence: 99%

An artificial neural network software tool for the assessment of the electric field around metal oxide surge arresters

Ekonomou

Mladenov

2015

Neural Comput & Applic

View full text Add to dashboard Cite

show abstract

“…The bottom three data sets deal with pattern classification and are used for comparing our methods with the NLPSVM (Newton method for Linear Programming SVM) [9]. (11) To examine that our proposed methods can delete redundant variables, we modify the Mackey-Glass data set [12], which is generated by a differential equation without noise. We add 18 artificial redundant input variables generated by a uniform random variable in [0,1] to the set.…”

Section: Benchmark Data Sets and Evaluation Conditionsmentioning

confidence: 99%

“…Boston 5 and Boston 14 data sets [13] use the 5th and 14th input variables of the Boston data set as outputs, respectively. Excluding the Mackey-Glass data set, these subject to (12) …”

Section: Benchmark Data Sets and Evaluation Conditionsmentioning

confidence: 99%

Fast Variable Selection by Block Addition and Block Deletion

Nagatani¹,

Ozawa²,

Abe

2010

JILSA

Self Cite

View full text Add to dashboard Cite

According to the selection criterion used, the variable selection methods are classified into wrapper methods and filter methods. Wrapper methods use the recognition rate as the selection criterion and filter methods use the selection criterion other than the recognition rate. Wrapper methods provide good generalization ability but usually at a much larger computational cost. Although the computational cost of the filter methods may be small, it will take a risk of selecting a subset of input variables that may deteriorate the generalization ability of the regressor.

show abstract

“…The water purification data set [7,8] is to estimate coagulant to be added to purify water. The Mackey-Glass data set [8,9] is a time series data set with chaotic behaviors. The Boston 5 and 14 data sets are from the Boston data set [10,11].…”

Section: Evaluation Conditionsmentioning

confidence: 99%

Support Vector Regression Using Mahalanobis Kernels

Kamada

Abe

2006

Artificial Neural Networks in Pattern Recognition

Self Cite

View full text Add to dashboard Cite

Abstract. In our previous work we have shown that Mahalanobis kernels are useful for support vector classifiers both from generalization ability and model selection speed. In this paper we propose using Mahalanobis kernels for function approximation. We determine the covariance matrix for the Mahalanobis kernel using all the training data. Model selection is done by line search. Namely, first the margin parameter and the error threshold are optimized and then the kernel parameter is optimized. According to the computer experiments for four benchmark problems, estimation performance of a Mahalanobis kernel with a diagonal covariance matrix optimized by line search is comparable to or better than that of an RBF kernel optimized by grid search.

show abstract

Neural Networks and Fuzzy Systems

Cited by 37 publications

References 13 publications

An artificial neural network software tool for the assessment of the electric field around metal oxide surge arresters

An artificial neural network software tool for the assessment of the electric field around metal oxide surge arresters

Fast Variable Selection by Block Addition and Block Deletion

Support Vector Regression Using Mahalanobis Kernels

Contact Info

Product

Resources

About