The fault feature extraction and classification of gear using principal component analysis and kernel principal component analysis based on the wavelet packet transform

Measurement ◽  
2014 ◽  
Vol 54 ◽  
pp. 118-132 ◽  
Author(s):  
Renping Shao ◽  
Wentao Hu ◽  
Yayun Wang ◽  
Xiankun Qi
Keyword(s):  
Principal Component Analysis ◽  
Feature Extraction ◽  
Wavelet Packet ◽  
Principal Component ◽  
Component Analysis ◽  
Wavelet Packet Transform ◽  
Kernel Principal Component Analysis ◽  
Fault Feature Extraction

Training Data Reduction and Classification Based on Greedy Kernel Principal Component Analysis and Fuzzy C-Means Algorithm

2013 ◽  
Vol 347-350 ◽  
pp. 2390-2394
Author(s):  
Xiao Fang Liu ◽  
Chun Yang
Keyword(s):  
Principal Component Analysis ◽  
Feature Extraction ◽  
Computational Complexity ◽  
Principal Component ◽  
Large Data ◽  
Component Analysis ◽  
Training Data ◽  
Kernel Principal Component Analysis ◽  
Nonlinear Feature Extraction ◽  
Nonlinear Feature

Nonlinear feature extraction used standard Kernel Principal Component Analysis (KPCA) method has large memories and high computational complexity in large datasets. A Greedy Kernel Principal Component Analysis (GKPCA) method is applied to reduce training data and deal with the nonlinear feature extraction problem for training data of large data in classification. First, a subset, which approximates to the original training data, is selected from the full training data using the greedy technique of the GKPCA method. Then, the feature extraction model is trained by the subset instead of the full training data. Finally, FCM algorithm classifies feature extraction data of the GKPCA, KPCA and PCA methods, respectively. The simulation results indicate that the feature extraction performance of both the GKPCA, and KPCA methods outperform the PCA method. In addition of retaining the performance of the KPCA method, the GKPCA method reduces computational complexity due to the reduced training set in classification.

Generation of critical aftershocks using stochastic neural networks and wavelet packet transform

2019 ◽  
Vol 26 (5-6) ◽  
pp. 331-351
Author(s):  
Elham Rajabi ◽  
Gholamreza Ghodrati Amiri
Keyword(s):  
Neural Networks ◽  
Principal Component Analysis ◽  
Response Spectrum ◽  
Wavelet Packet ◽  
Principal Component ◽  
Response Spectra ◽  
Component Analysis ◽  
Wavelet Packet Transform ◽  
Stochastic Neural Networks ◽  
Transform Coefficients

This paper proposes a methodology using wavelet packet transform, principal component analysis, and neural networks in order to generate artificial critical aftershock accelerograms which are compatible with the response spectra. This procedure uses the learning abilities of neural networks, principal component analysis as a dimension reduction technique, and decomposing capabilities of wavelet packet transform on consecutive earthquakes. In fact, the proposed methodology consists of two steps and expands the knowledge of the inverse mapping from mainshock response spectrum to aftershock response spectrum and aftershock response spectrum to wavelet packet transform coefficients of the aftershocks. This procedure results in a stochastic ensemble of response spectra of aftershock (first step) and corresponding wavelet packet transform coefficients (second step) which are then used to generate the aftershocks through applying the inverse wavelet packet transform. Finally, in order to demonstrate the effectiveness of the proposed method, three examples are presented in which recorded critical successive ground motions are used to train and test the neural networks.

Laser-Welding Spots Detection Based on Original Moment Values

2014 ◽  
Vol 599-601 ◽  
pp. 974-980
Author(s):  
Xiao Long Qi ◽  
Bin Fang ◽  
Shu Mei Wang
Keyword(s):  
Principal Component Analysis ◽  
Laser Welding ◽  
Image Classification ◽  
Principal Component ◽  
Component Analysis ◽  
Kernel Principal Component Analysis ◽  
Invariant Moments ◽  
The Past

In the past decades, the theories of invariant moments have been researched extensively and wildly used in many fields. However, for the laser-welding spots of titanium tubes or other fixed objects, the invariant moments are inapplicable. Besides, the studies and experiments about image classification by means of the original moment values were barely proposed. In this paper, the method of classification based on original moment values is introduced, and an improved approach of KPCA (kernel principal component analysis) in order to reduce the inner-class distance of the qualified laser-welding spots is also discussed. Finally, experiments are carried out to validate the classification ability, and results show that the original moment values are suited as pattern features in classification of fixed objects.

Sign in / Sign up

Export Citation Format

Share Document