Feature selection for noisy variation patterns using kernel principal component analysis

Sahu, Anshuman Kumar; Apley, Daniel W.; Runger, George C.

doi:10.1016/j.knosys.2014.08.027

Cited by 18 publications

(10 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Another type of nonlinear dimensionality reduction method is the kernel-based methods. With the successful application of kernel methods in the support vector machines, kernel-based dimensionality reduction methods, such as Kernel principal component Analysis (KPCA) [25], Kernel discriminant analysis (KDA) [26], Generalized discriminant analysis (GDA) [27], and Kernel Canonical Correlation Analysis (KCCA) [28], are widely used in nonlinear pattern classification. However, the form of kernel function determines the geometric structure of feature space, and has a crucial impact on the performance of kernel based methods.…”

Section: Introductionmentioning

confidence: 99%

Multiple Kernel-Based Discriminant Analysis via Support Vectors for Dimension Reduction

et al. 2019

View full text Add to dashboard Cite

Kernel-based discriminant analysis is an effective nonlinear mechanism for pattern analysis. Conventional kernel-based discriminant analysis mainly based on a single kernel function may be insufficient when dealing with datasets with complicated geometric structures. A combination of multiple kernels is able to represent the complementary information of the original data from multiple views and thereby improves recognition performance. However, the discriminant analysis methods based on the combination of multiple kernels face the challenges of optimizing the weights of the ''base kernels'' and the heavy computational burden. To address these challenges, this paper proposes a novel multi-kernel discriminant analysis method based on support vectors (MKDASV) to represent the data structure more effectively by incorporating the between-class and within-class information. First, the multi-kernel SVM algorithm is utilized to obtain the weight of each ''base kernel'' and the support vectors; and then the criteria of discriminant analysis method are constructed by taking into account both the margin maximizing classification theory of SVM and the expression of the within-class scatter in LDA algorithm; and finally, to effectively reduce the amount of computation, only the support vectors are used as the training samples to participate in the dimensionality reduction operation. The experimental results on six standard databases validated that our proposed method outperformed the other five methods in terms of classification accuracy and the computational efficiency as well. INDEX TERMS Kernel discriminant analysis, multi-kernel SVM, support vectors.

show abstract

Section: Introductionmentioning

confidence: 99%

Multiple Kernel-Based Discriminant Analysis via Support Vectors for Dimension Reduction

et al. 2019

View full text Add to dashboard Cite

show abstract

“…In fault diagnosis, we first use these methods to obtain the original signal features. Then, we use other methods, such as genetic algorithm [11][12][13][14], principal component analysis [15,16] and kernel principal component analysis [17,18] to perform necessary selections and transformations to obtain the appropriate features.…”

Section: Introductionmentioning

confidence: 99%

Method for the extraction of shock signal features based on the upper limit of density integral

Yang¹,

Hong-xia²

2019

J. vibroeng.

View full text Add to dashboard Cite

Shock signal features must be extracted for use in pattern recognition or fault diagnosis. In this work, we proposed a method for the feature extraction of shock signals, which are vibration signals that change faster and have larger amplitude ranges than general signals. First, we proposed the concepts of amplitude density for monotonic functions and piecewise monotonic functions. On the basis of these concepts, we then proposed the concept of the upper limit of density integral (ULDI), which was adopted to obtain signal features. Then, we introduced two types of serious fault cracks to the latch sheet of an automatic gun mechanism that is used on warships. Next, we applied the proposed method to extract the features of shock signals from data acquired when the automatic gun mechanism fired with normal and two fault patterns. Finally, we verified the effectiveness of our proposed method by applying the features that it extracted to a support vector machine (SVM). Our proposed method provided good results and was superior to the traditional statistics-based feature extraction method when applied to a SVM for classification. In addition, the former method demonstrated better generalisation than the latter. Thus, our method is an efficient approach for extracting shock signal features in pattern recognition and fault diagnosis.

show abstract

“…It is known that familiar pre-processing methods of dimension reduction include principal component analysis (PCA), linear discriminant analysis, and locally linear embedding (LLE), among others. Sahu et al (2014) proposed a feature selection procedure that augmented kernel PCA to obtain importance estimates of the features using noisy training data. However, PCA can only identify linear relationships among features in the data.…”

Section: Introductionmentioning

confidence: 99%

Locally linear embedding and neighborhood rough set-based gene selection for gene expression data classification

Sun

Wang

et al. 2016

Genet. Mol. Res.

View full text Add to dashboard Cite

ABSTRACT. Cancer subtype recognition and feature selection are important problems in the diagnosis and treatment of tumors. Here, we propose a novel gene selection approach applied to gene expression data classification. First, two classical feature reduction methods including locally linear embedding (LLE) and rough set (RS) are summarized. The advantages and disadvantages of these algorithms were analyzed and an optimized model for tumor gene selection was developed based on LLE and neighborhood RS (NRS). Bhattacharyya distance was introduced to delete irrelevant genes, pair-wise redundant analysis was performed to remove strongly correlated genes, and the wavelet soft threshold was determined to eliminate noise in the gene datasets. Next, prior optimized search processing was carried out. A new approach combining dimension reduction of LLE and feature reduction of NRS (LLE-NRS) was developed for selecting gene subsets, and then an open source software Weka was applied to distinguish different tumor types and verify the cross-validation classification accuracy of our proposed method. The experimental results demonstrated that the classification performance of the proposed LLE-NRS for selecting gene subset outperforms those of other related models in terms of accuracy, and our proposed approach is feasible and effective in the field of highdimensional tumor classification.

show abstract

Feature selection for noisy variation patterns using kernel principal component analysis

Cited by 18 publications

References 22 publications

Multiple Kernel-Based Discriminant Analysis via Support Vectors for Dimension Reduction

Multiple Kernel-Based Discriminant Analysis via Support Vectors for Dimension Reduction

Method for the extraction of shock signal features based on the upper limit of density integral

Locally linear embedding and neighborhood rough set-based gene selection for gene expression data classification

Contact Info

Product

Resources

About