DocumentCode :
1332792
Title :
Robust Single-Hidden Layer Feedforward Network-Based Pattern Classifier
Author :
Zhihong Man ; Lee, Kahyun ; Dianhui Wang ; Zhenwei Cao ; Suiyang Khoo
Author_Institution :
Fac. of Eng. & Ind. Sci., Swinburne Univ. of Technol., Melbourne, VIC, Australia
Volume :
23
Issue :
12
fYear :
2012
Firstpage :
1974
Lastpage :
1986
Abstract :
In this paper, a new robust single-hidden layer feedforward network (SLFN)-based pattern classifier is developed. It is shown that the frequency spectrums of the desired feature vectors can be specified in terms of the discrete Fourier transform (DFT) technique. The input weights of the SLFN are then optimized with the regularization theory such that the error between the frequency components of the desired feature vectors and the ones of the feature vectors extracted from the outputs of the hidden layer is minimized. For the linearly separable input patterns, the hidden layer of the SLFN plays the role of removing the effects of the disturbance from the noisy input data and providing the linearly separable feature vectors for the accurate classification. However, for the nonlinearly separable input patterns, the hidden layer is capable of assigning the DFTs of all feature vectors to the desired positions in the frequency-domain such that the separability of all nonlinearly separable patterns are maximized. In addition, the output weights of the SLFN are also optimally designed so that both the empirical and the structural risks are well balanced and minimized in a noisy environment. Two simulation examples are presented to show the excellent performance and effectiveness of the proposed classification scheme.
Keywords :
discrete Fourier transforms; feature extraction; feedforward neural nets; image classification; optimisation; DFT; SLFN-based pattern classifier; discrete Fourier transform; feature vector extraction; feature vector minimization; frequency components; frequency spectrums; input weight optimization; linearly separable feature vectors; linearly separable input patterns; noisy input data; nonlinearly separable pattern separability maximisation; regularization theory; robust single-hidden layer feedforward network; Discrete Fourier transforms; Feature extraction; Feedforward neural networks; Frequency domain analysis; Robustness; Training; Vectors; Discrete Fourier transform; feedforward networks; pattern classification; regularization theory;
fLanguage :
English
Journal_Title :
Neural Networks and Learning Systems, IEEE Transactions on
Publisher :
ieee
ISSN :
2162-237X
Type :
jour
DOI :
10.1109/TNNLS.2012.2218616
Filename :
6352923
Link To Document :
بازگشت