Weighted Kernel Principal Component Analysis Research Articles

Fault diagnosis of transformer winding short circuit based on WKPCA-WM and IPOA-CNN

When the winding of the power transformer is short-circuited, the winding will experience constant vibration, which will cause axial instability of the winding, and then lead to winding looseness, deformation, bulge, etc., therefore, a diagnosis method based on the Improved Pelican Optimization Algorithm and Convolutional Neural Network (IPOA-CNN) for short-circuit voiceprint signal of transformer windings is proposed. At the same time, considering the input parameter dimension of deep learning cannot be too high, a new feature parameter selection method is constructed for this model. Firstly, the frequency characteristics of winding acoustic vibration signals are analyzed, and then the characteristic parameters of transformer acoustic signals are extracted by Wavelet Packet Energy Spectrum (WPES) and Mel Frequency Cepstrum Coefficient (MFCC), respectively. Then, the two methods are combined to construct the WM feature extraction algorithm, and the Weighted Kernel Principal Component Analysis (WKPCA) is used to reduce the dimension of the feature to obtain the feature parameters with accurate feature information and low redundancy; Finally, combined with Sobol sequence to optimize the initial population of Pelican Optimization Algorithm (POA), the convolution kernel of Convolutional neural network (CNN) was optimized by IPOA, and the optimal convolution kernel was obtained. The transformer winding short-circuits voiceprint diagnosis models of WKPCA-WM and IPOA-CNN were constructed, which realized the accurate diagnosis of winding short-circuit voiceprint. The validity and feasibility of the method are verified by the acoustic signal data collected in the laboratory.

Research on Multi-Fault Identification of Marine Vertical Centrifugal Pump Based on Multi-Domain Characteristic Parameters

The marine vertical centrifugal pump is an important piece of auxiliary equipment for ships. Due to the complex operating conditions of marine equipment and the frequent swaying of the hull, typical pump failures such as rotor misalignment, rotor unbalance and mechanical loosening occur frequently, which seriously affect the service life of the marine vertical centrifugal pump. Based on multi-domain characteristic parameters, a fault identification method combining weighted kernel principal component analysis (WKPCA) and particle swarm optimization support vector machine (PSO-SVM) is proposed in this paper. It can effectively solve the problem of multi-fault classification of the centrifugal pump and provide reference for efficient maintenance of equipment. Firstly, a vertical centrifugal pump test bench is set up to simulate typical faults. The collected original fault data are denoised by Kalman filtering. Then, a multi-domain feature set composed of 20 feature parameters was constructed. However, due to high dimension, data redundancy and calculation time were increased. After dimensionality reduction, a fault feature set with 9 feature indexes was established by combining with the WKPCA method. Finally, the PSO-SVM model is used to realize multi-fault identification, and the recognition results of the traditional support vector machine and the genetic algorithm support vector machine (GA-SVM) are compared to verify the diagnosis results and classification performance of PSO-SVM. The results show that the accuracy of WKPCA and PSO-SVM fault recognition methods based on multi-domain characteristic parameters is 1, and it has good convergence.

Method of fault feature selection and fusion based on poll mode and optimized weighted KPCA for bearings

The health monitoring system for equipment is essential in the smooth proceeding of industrial production. However, the fault features to be detected in monitoring systems are generally selected through projects and expertise, which are not capable for complex and ever-changing fault information and may result in incomplete correspondence to the fault types emerged. To dig deeper for the effective features hidden in the data instead of selecting by experience, a feature selection and fusion method based on poll mode and optimized Weighted Kernel Principal Component Analysis (WKPCA) method is then proposed. Specifically, inspired by poll-mode and multi-criteria strategy, a multi-measure hierarchical model is designed to sort the fault features with high sensitivity, acquiring the feature subset with corresponding weight coefficient. Considering the variation in fault information collected by different sensors, the diagnosis rate in Extreme Learning Machine (ELM) is taken as the index for evaluation of each single sensor, then the sensitivity weight matrix of features extracted by multiple sensors is constructed after linear normalization. To integrate the feature information, WKPCA is applied for the weighted fusion of features, and Quantum Genetic Algorithm (QGA) is used to search for the kernel width parameter when the best separability in the samples under the fusion is reached. Finally, such samples are introduced to drive the diagnostic model of the monitoring system in rolling bearing. The experimental results show that, compared with the traditional feature selection and fusion methods, this method is capable for sorting out highly sensitive features with more fault information self-adaptively, and can improve the separability in the subset of fault samples effectively.

A novel dimension reduction algorithm based on weighted kernel principal analysis for gene expression data.

Gene expression data has the characteristics of high dimensionality and a small sample size and contains a large number of redundant genes unrelated to a disease. The direct application of machine learning to classify this type of data will not only incur a great time cost but will also sometimes fail to improved classification performance. To counter this problem, this paper proposes a dimension-reduction algorithm based on weighted kernel principal component analysis (WKPCA), constructs kernel function weights according to kernel matrix eigenvalues, and combines multiple kernel functions to reduce the feature dimensions. To further improve the dimensional reduction efficiency of WKPCA, t-class kernel functions are constructed, and corresponding theoretical proofs are given. Moreover, the cumulative optimal performance rate is constructed to measure the overall performance of WKPCA combined with machine learning algorithms. Naive Bayes, K-nearest neighbour, random forest, iterative random forest and support vector machine approaches are used in classifiers to analyse 6 real gene expression dataset. Compared with the all-variable model, linear principal component dimension reduction and single kernel function dimension reduction, the results show that the classification performance of the 5 machine learning methods mentioned above can be improved effectively by WKPCA dimension reduction.

Kernel-based framework for spectral dimensionality reduction and clustering formulation: A theoretical study

This work outlines a unified formulation to represent spectral approaches for both dimensionality reduction and clustering. Proposed formulation starts with a generic latent variable model in terms of the projected input data matrix.Particularly, such a projection maps data onto a unknown high-dimensional space. Regarding this model, a generalized optimization problem is stated using quadratic formulations and a least-squares support vector machine.The solution of the optimization is addressed through a primal-dual scheme.Once latent variables and parameters are determined, the resultant model outputs a versatile projected matrix able to represent data in a low-dimensional space, as well as to provide information about clusters. Particularly, proposedformulation yields solutions for kernel spectral clustering and weighted-kernel principal component analysis.

Registration for Variform Object of Remote-Sensing Image Using Improved Robust Weighted Kernel Principal Component Analysis

For pre- and post-earthquake remote-sensing images, registration is a challenging task due to the possible deformations of the objects to be registered. To overcome this problem, in a previous paper, we proposed a registration method based on robust weighted kernel principal component analysis (RWKPCA) to precisely register the variform objects. Which was proved very effective in capturing the common robust kernel principal components (RKPCs) and generalized well for registration. Compared with previous paper, there are two improvements in this paper: Firstly, we developed the improved RWKPCA method from the robust loss function, and theoretically proved the robustness of the method; Secondly, a new construction of weight function by projection residual was given, which enables the great reduction of computing time. Finally, two experiments were conducted on the remote-sensing image registration in Wenchuan earthquake and change detection of Tangjiashan barrier lake, and the results showed that compared with the previous method, the registration accuracy was increased while the computational time was decreased a lot. Meanwhile, good performance on the change detection of barrier lake is obtained.

Fault Detection of a Proposed Three-Level Inverter Based on a Weighted Kernel Principal Component Analysis

Fault detection is the research focus and priority in this study to ensure the high reliability of a proposed three-level inverter. Kernel principal component analysis (KPCA) has been widely used for feature extraction because of its simplicity. However, highlighting useful information that may be hidden under retained KPCs remains a problem. A weighted KPCA is proposed to overcome this shortcoming. Variable contribution plots are constructed to evaluate the importance of each KPC on the basis of sensitivity analysis theory. Then, different weighting values of KPCs are set to highlight the useful information. The weighted statistics are evaluated comprehensively by using the improved feature eigenvectors. The effectiveness of the proposed method is validated. The diagnosis results of the inverter indicate that the proposed method is superior to conventional KPCA.

Multilevel hierarchical kernel spectral clustering for real-life large scale complex networks.

Kernel spectral clustering corresponds to a weighted kernel principal component analysis problem in a constrained optimization framework. The primal formulation leads to an eigen-decomposition of a centered Laplacian matrix at the dual level. The dual formulation allows to build a model on a representative subgraph of the large scale network in the training phase and the model parameters are estimated in the validation stage. The KSC model has a powerful out-of-sample extension property which allows cluster affiliation for the unseen nodes of the big data network. In this paper we exploit the structure of the projections in the eigenspace during the validation stage to automatically determine a set of increasing distance thresholds. We use these distance thresholds in the test phase to obtain multiple levels of hierarchy for the large scale network. The hierarchical structure in the network is determined in a bottom-up fashion. We empirically showcase that real-world networks have multilevel hierarchical organization which cannot be detected efficiently by several state-of-the-art large scale hierarchical community detection techniques like the Louvain, OSLOM and Infomap methods. We show that a major advantage of our proposed approach is the ability to locate good quality clusters at both the finer and coarser levels of hierarchy using internal cluster quality metrics on 7 real-life networks.

Weighted KPCA Degree of Homogeneity Amended Nonclassical Receptive Field Inhibition Model for Salient Contour Extraction in Low-Light-Level Image

The stimulus response of the classical receptive field (CRF) of neuron in primary visual cortex is affected by its periphery [i.e., non-CRF (nCRF)]. This modulation exerts inhibition, which depends primarily on the correlation of both visual stimulations. The theory of periphery and center interaction with visual characteristics can be applied in night vision information processing. In this paper, a weighted kernel principal component analysis (WKPCA) degree of homogeneity (DH) amended inhibition model inspired by visual perceptual mechanisms is proposed to extract salient contour from complex natural scene in low-light-level image. The core idea is that multifeature analysis can recognize the homogeneity in modulation coverage effectively. Computationally, a novel WKPCA algorithm is presented to eliminate outliers and anomalous distribution in CRF and accomplish principal component analysis precisely. On this basis, a new concept and computational procedure for DH is defined to evaluate the dissimilarity between periphery and center comprehensively. Through amending the inhibition from nCRF to CRF by DH, our model can reduce the interference of noises, suppress details, and textures in homogeneous regions accurately. It helps to further avoid mutual suppression among inhomogeneous regions and contour elements. This paper provides an improved computational visual model with high-performance for contour detection from cluttered natural scene in night vision image.

Feature extraction of wound infection data for electronic nose based on a novel weighted KPCA

When an electronic nose (E-nose) is used to predict the classes of wound infection, its result is not ideal if the original feature matrix extracted from the response of sensors is put into the classifier directly. To acquire more useful information which can improve E-nose's classification accuracy, we present a novel weighted kernel principal component analysis (KPCA) method to process this matrix. In addition, we have also compared it with other existing methods including independent component analysis (ICA), orthogonal signal correction (OSC), locality preserving projections (LPP), principal component analysis (PCA), KPCA and the traditional weighted KPCA. The odors of four different classes of wounds (uninfected and infected with Staphylococcus aureus, Escherichia coli, and Pseudomonas aeruginosa) are used as the original response of E-nose. Experimental results have demonstrated that the proposed weighted KPCA method outperforms other feature extraction methods.

Weighted kernel principal component analysis based on probability density estimation and moving window and its application in nonlinear chemical process monitoring

Kernel principal component analysis (KPCA) has been widely used in nonlinear process monitoring; however, KPCA does not always perform efficiently because useful information may be submerged under retained KPCs. To address this shortcoming, probability density estimation- and moving weighted window-based KPCA (PM-WKPCA) is proposed. PM-WKPCA is used mainly to estimate the probability and evaluate the importance of each KPC by kernel density estimation and then set different weighting values on KPCs to highlight the useful information. The status of the process is also evaluated comprehensively using weighted statistics within a moving window. The efficiency of the proposed method is demonstrated by the following: case studies on a numerical nonlinear system, the simulated continuously stirred tank reactor process, and the Tennessee Eastman process. Monitoring results indicate that the proposed method is superior to the conventional PCA, KPCA, and some typical extension methods.

Adaptive multiparameter spectral feature analysis for synthetic aperture radar target recognition

A feature extraction algorithm based on spectral clustering with adaptive multiparameters is proposed for synthetic aperture radar automatic target recognition (SAR-ATR). Spectral clustering has been widely applied in computer vision for its good performance. Meanwhile, the spectral mapping step in it has the property of feature space transformation. Spectral clustering based target feature extraction for SAR-ATR is constructed according to the framework of out-of-sample extensions in weighted kernel principal component analysis. To avoid the scaling parameter selection in spectral feature analysis (SFA) and eliminate the influence of scaling parameter on feature extraction performance as well, the multiple scaling parameters are calculated adaptively by local neighborhoods. Because the local statistics of the neighborhood of each point are taken into consideration, its performance is better than using only one fixed parameter. Based on the extracted features, target recognition is performed by the support vector machine for its good generalization capability. The experimental results show that the multiparameter SFA outperforms the principal component analysis, kernel principal component analysis and SFA with the selected scaling parameter for SAR target recognition in terms of recognition accuracy.

Registration of remote-sensing images using robust weighted kernel principal component analysis

Abstract For pre- and post-earthquake remote-sensing images, registration is a challenging task due to the possible deformations of the objects to be registered. To overcome this problem, a registration method based on robust weighted kernel principal component analysis is proposed to precisely register the variform objects. Firstly, a robust weighted kernel principal component analysis (RWKPCA) method is developed to capture the common robust kernel principal components (RKPCs) of the variform objects. Secondly, a registration approach is derived from the projection on RKPCs. Finally, two experiments are conducted on the SAR image registration in Wenchuan earthquake on May 12, 2008, and the results showed that the method is very effective in capturing structure patterns and generalized well for registration.

Kernel spectral clustering of time series in the CoRoT exoplanet database

Context. Detection of contaminated light curves and irregular variables has become a challenge when studying variable stars in large photometric surveys such as that produced by the CoRoT mission.Aims. Our goal is to characterize and cluster the light curves of the first four runs of CoRoT, in order to find the stars that cannot be classified because of either contamination or exceptional or non-periodic behavior.Methods. We study three different approaches to characterize the light curves, namely Fourier parameters, autocorrelation functions (ACF), and hidden Markov models (HMMs). Once the light curves have been transformed into a different input space, they are clustered, using kernel spectral clustering. This is an unsupervised technique based on weighted kernel principal component analysis (PCA) and least squares support vector machine (LS-SVM) formulations. The results are evaluated using the silhouette value.Results. The most accurate characterization of the light curves is obtained by means of HMM. This approach leads to the identification of highly contaminated light curves. After kernel spectral clustering has been implemented onto this new characterization, it is possible to separate the highly contaminated light curves from the rest of the variables. We improve the classification of binary systems and identify some clusters that contain irregular variables. A comparison with supervised classification methods is also presented.

Multiway Spectral Clustering with Out-of-Sample Extensions through Weighted Kernel PCA

A new formulation for multiway spectral clustering is proposed. This method corresponds to a weighted kernel principal component analysis (PCA) approach based on primal-dual least-squares support vector machine (LS-SVM) formulations. The formulation allows the extension to out-of-sample points. In this way, the proposed clustering model can be trained, validated, and tested. The clustering information is contained on the eigendecomposition of a modified similarity matrix derived from the data. This eigenvalue problem corresponds to the dual solution of a primal optimization problem formulated in a high-dimensional feature space. A model selection criterion called the Balanced Line Fit (BLF) is also proposed. This criterion is based on the out-of-sample extension and exploits the structure of the eigenvectors and the corresponding projections when the clusters are well formed. The BLF criterion can be used to obtain clustering parameters in a learning framework. Experimental results with difficult toy problems and image segmentation show improved performance in terms of generalization to new samples and computation times.