scispace - formally typeset
Search or ask a question

Showing papers on "Mahalanobis distance published in 2019"


Journal ArticleDOI
TL;DR: The proposed cascade classification framework consistently yields a high performance for retinal vessel segmentation, and delineates a more complete and accurate vessel tree, and can be flexibly extended to other image recognition tasks.

97 citations


Journal ArticleDOI
19 Feb 2019-Water SA
TL;DR: The aim of this study was to compare the classification of selected vegetation types using both hyperspectral and multispectral satellite remote sensing data.
Abstract: In recent years the use of remote sensing imagery to classify and map vegetation over different spatial scales has gained wide acceptance in the research community. Many national and regional datasets have been derived using remote sensing data. However, much of this research was undertaken using multispectral remote sensing datasets. With advances in remote sensing technologies, the use of hyperspectral sensors which produce data at a higher spectral resolution is being investigated. The aim of this study was to compare the classification of selected vegetation types using both hyperspectral and multispectral satellite remote sensing data. Several statistical classifiers including maximum likelihood, minimum distance, mahalanobis distance, spectral angular mapper and parallelepiped methods of classification were used. Classification using mahalanobis distance and maximum likelihood methods with an optimal set of hyperspectral and multispectral bands produced overall accuracies greater than 80%.

83 citations


Journal ArticleDOI
TL;DR: This paper introduces a suppression function to construct a discriminative feature space and utilizes a deep brief network to learn spectral representation and abstraction automatically that are used as inputs to the Mahalanobis distance (MD)-based detector.
Abstract: Hyperspectral anomaly detection faces various levels of difficulty due to the high dimensionality of hyperspectral images (HSIs), redundant information, noisy bands, and the limited capability of utilizing spectral–spatial information. In this paper, we address these problems and propose a novel approach, called spectral–spatial feature extraction (SSFE), which is based on two main aspects. In the spectral domain, we assume that the anomalous pixels are rarely present and all (or most) of the samples around the anomalies belong to background (BKG). Using this fact, we introduce a suppression function to construct a discriminative feature space and utilize a deep brief network to learn spectral representation and abstraction automatically that are used as inputs to the Mahalanobis distance (MD)-based detector. In the spatial domain, the anomalies appear as a small area grouped by pixels with high correlation among them compared to BKG. Therefore, the objects appearing as a small area are extracted based on attribute filtering, and a guided filter is further employed for local smoothness. More specifically, we extract spatial features of anomalies only from one single band obtained by fusing all bands in the visible wavelength range. Finally, we detect anomalies by jointly considering the spectral and spatial detection results. Several experiments are performed, which show that our proposed method outperforms the state-of-the-art methods.

67 citations


Posted Content
TL;DR: The hypothesis that a simple class-covariance-based distance metric, namely the Mahalanobis distance, adopted into a state of the art few-shot learning approach (CNAPS) can, in and of itself, lead to a significant performance improvement is explored.
Abstract: Few-shot learning is a fundamental task in computer vision that carries the promise of alleviating the need for exhaustively labeled data. Most few-shot learning approaches to date have focused on progressively more complex neural feature extractors and classifier adaptation strategies, as well as the refinement of the task definition itself. In this paper, we explore the hypothesis that a simple class-covariance-based distance metric, namely the Mahalanobis distance, adopted into a state of the art few-shot learning approach (CNAPS) can, in and of itself, lead to a significant performance improvement. We also discover that it is possible to learn adaptive feature extractors that allow useful estimation of the high dimensional feature covariances required by this metric from surprisingly few samples. The result of our work is a new "Simple CNAPS" architecture which has up to 9.2% fewer trainable parameters than CNAPS and performs up to 6.1% better than state of the art on the standard few-shot image classification benchmark dataset.

57 citations


Journal ArticleDOI
04 Oct 2019
TL;DR: In this paper, after short reviewing some tools for univariate outliers detection, the Mahalanobis distance, as a famous multivariate statistical distances, and its ability to detect multivariate outsiers are discussed.
Abstract: While methods of detecting outliers is frequently implemented by statisticians when analyzing univariate data, identifying outliers in multivariate data pose challenges that univariate data do not. In this paper, after short reviewing some tools for univariate outliers detection, the Mahalanobis distance, as a famous multivariate statistical distances, and its ability to detect multivariate outliers are discussed. As an application the univariate and multivariate outliers of a real data set has been detected using R software environment for statistical computing.

57 citations


Journal ArticleDOI
TL;DR: In this article, a new methodology based on PLS regression models is proposed considering a reparameterized Birnbaum-Saunders (RBS) distribution for the response, which is useful for describing asymmetric data frequently found in chemical phenomena.

47 citations


Posted Content
TL;DR: This tutorial explains Linear Discriminant Analysis and Quadratic Discriminatory Analysis as two fundamental classification methods in statistical and probabilistic learning and proves that LDA and Fisher discriminant analysis are equivalent.
Abstract: This tutorial explains Linear Discriminant Analysis (LDA) and Quadratic Discriminant Analysis (QDA) as two fundamental classification methods in statistical and probabilistic learning. We start with the optimization of decision boundary on which the posteriors are equal. Then, LDA and QDA are derived for binary and multiple classes. The estimation of parameters in LDA and QDA are also covered. Then, we explain how LDA and QDA are related to metric learning, kernel principal component analysis, Mahalanobis distance, logistic regression, Bayes optimal classifier, Gaussian naive Bayes, and likelihood ratio test. We also prove that LDA and Fisher discriminant analysis are equivalent. We finally clarify some of the theoretical concepts with simulations we provide.

40 citations


Journal ArticleDOI
TL;DR: A Mahalanobis distance–based Monte Carlo goodness of fit testing procedure for the family of stochastic actor-oriented models for social network evolution and a modified model distance estimator is proposed to help the researcher identify model extensions that will remediate poor fit.
Abstract: We propose a Mahalanobis distance–based Monte Carlo goodness of fit testing procedure for the family of stochastic actor-oriented models for social network evolution. A modified model distance esti...

39 citations


Journal ArticleDOI
TL;DR: A Semi-Supervised Metric Transfer Learning framework called SSMT is proposed that reduces the distribution between domains both statistically and geometrically by learning the instance weights, while a regularized distance metric is learned to minimize the within- class co-variance and maximize the between-class co-Variance simultaneously for the target domain.
Abstract: A common assumption of statistical learning theory is that the training and testing data are drawn from the same distribution However, in many real-world applications, this assumption does not hold true Hence, a realistic strategy, Cross Domain Adaptation (DA) or Transfer Learning (TA), can be used to employ previously labelled source domain data to boost the task in the new target domain Previously, Cross Domain Adaptation methods have been focused on re-weighting the instances or aligning the cross-domain distributions However, these methods have two significant challenges: (1) There is no proper consideration of the unlabelled data of target task as in the real-world, an abundant amount of unlabelled data is available, (2) The use of normal Euclidean distance function fails to capture the appropriate similarity or dissimilarity between samples To deal with this issue, we have proposed a Semi-Supervised Metric Transfer Learning framework called SSMT that reduces the distribution between domains both statistically and geometrically by learning the instance weights, while a regularized distance metric is learned to minimize the within-class co-variance and maximize the between-class co-variance simultaneously for the target domain Compared with the previous works where Mahalanobis distance metric and instance weights are learned by using the labelled data or in a pipelined framework that leads to a decrease in the performance, our proposed SSMT attempts to learn a regularized distance metric and instance weights by considering unlabelled data in a parallel framework Experimental evaluation on three cross-domain visual data sets, eg, PIE Face, Handwriting Digit Recognition on MNIST–USPS and Object Recognition, demonstrates the effectiveness of our designed approach on facilitating the unlabelled target task learning, compared to current state-of-the-art domain adaptation approaches

38 citations


Journal ArticleDOI
TL;DR: A faster convergence rate of DML is obtained, O1N, when learning the distance metric with a smooth loss function and a strongly convex objective, and when the problem is relatively easy, and the number of training samples is large enough, this rate can be further improved.
Abstract: Distance metric learning (DML) aims to find a suitable measure to compute a distance between instances. Facilitated by side information, the learned metric can often improve the performance of similarity or distance based methods such as kNN. Theoretical analyses of DML focus on the learning effectiveness for squared Mahalanobis distance. Specifically, whether the Mahalanobis metric learned from the empirically sampled pairwise constraints is in accordance with the optimal metric optimized over the paired samples generated from the true distribution, and the sample complexity of this process. The excess risk could measure the quality of the generalization, i.e., the gap between the expected objective of empirical metric learned from a regularized objective with convex loss function and the one with the optimal metric. Given N training examples, existing analyses over this non-i.i.d. learning problem have proved the excess risk of DML converges to zero at a rate of $${\mathcal {O}}\left( \frac{1}{\sqrt{N}}\right) $$ . In this paper, we obtain a faster convergence rate of DML, $${\mathcal {O}}\left( \frac{1}{N}\right) $$ , when learning the distance metric with a smooth loss function and a strongly convex objective. In addition, when the problem is relatively easy, and the number of training samples is large enough, this rate can be further improved to $${\mathcal {O}}\left( \frac{1}{N^2}\right) $$ . Synthetic experiments validate that DML can achieve the specified faster generalization rate, and results under various settings help explore the theoretical properties of DML a lot.

36 citations


Journal ArticleDOI
TL;DR: With the proposed method, health monitoring of rotating machinery could be achieved without prior information and domain knowledge, thereby providing an automatic data processing and condition monitoring tool in big data context.
Abstract: In the era of big data, a huge amount of monitoring and manufacturing data is generated every hour. As these data are typically measured from different machines and under different working regimes, prior information and domain knowledge are highly required in order to properly analyze and utilize these data. In view of this limitation, a data-driven self-comparison approach is proposed for the monitoring of rotating machinery. In this approach, comb filtering is introduced to extract the concerned signals from multisource background noise. A Gini-guided residual singular value decomposition is then proposed to enhance local anomalies induced by early defects. Finally, an iterative Mahalanobis distance is constructed to measure the statistical deviation of monitored component from a normal state. With the proposed method, health monitoring of rotating machinery could be achieved without prior information and domain knowledge, thereby providing an automatic data processing and condition monitoring tool in big data context.

Journal ArticleDOI
TL;DR: This article focuses on the performance of different estimations of the Mahalanobis distance metric using robust estimates for location and scatter, and these alternative formulations are compared to traditional, less robust estimation methods.
Abstract: In this study, a damage detection and localization scenario is presented for a composite laminate with a network of embedded fiber Bragg gratings. Strain time histories from a pseudorandom simulated operational loading are mined for multivariate damage-sensitive feature vectors that are then mapped to the Mahalanobis distance, a covariance-weighted distance metric for discrimination. The experimental setup, data acquisition, and feature extraction are discussed briefly, and special attention is given to the statistical model used for a binary hypothesis test for damage diagnosis. This article focuses on the performance of different estimations of the Mahalanobis distance metric using robust estimates for location and scatter, and these alternative formulations are compared to traditional, less robust estimation methods.

Journal ArticleDOI
02 Apr 2019-PeerJ
TL;DR: How Mahalanobis distances are calculated is explained, and how to correctly produce probabilities is demonstrated, to maximise the potential application of the MahalanOBis distance technique within the ecological modelling community.
Abstract: The Mahalanobis distance is a statistical technique that can be used to measure how distant a point is from the centre of a multivariate normal distribution By measuring Mahalanobis distances in environmental space ecologists have also used the technique to model: ecological niches, habitat suitability, species distributions, and resource selection functions Unfortunately, the original description of the Mahalanobis distance technique for ecological modelling contained an error describing how Mahalanobis distances could be converted into probabilities using a chi-squared distribution This error has been repeated in the literature, and is present in popular modelling software In the hope of correcting this error to maximise the potential application of the Mahalanobis distance technique within the ecological modelling community, I explain how Mahalanobis distances are calculated, and through a virtual ecology experiment demonstrate how to correctly produce probabilities and discuss the implications of the error for previous Mahalanobis distance studies

Journal ArticleDOI
TL;DR: This report shows how differences in the variance of features lead to differences inThe strength of the influence of each feature on the similarity scores produced from all the features, and compares six variance normalization methods in terms of how well they reduce the impact of the variance differences.
Abstract: The importance of normalizing biometric features or matching scores is understood in the multimodal biometric case, but there is less attention to the unimodal case. Prior reports assess the effectiveness of normalization directly on biometric performance. We propose that this process is logically comprised of two independent steps: (1) methods to equalize the effect of each biometric feature on the similarity scores calculated from all the features together and (2) methods of weighting the normalized features to optimize biometric performance. In this report, we address step 1 only and focus exclusively on normally distributed features. We show how differences in the variance of features lead to differences in the strength of the influence of each feature on the similarity scores produced from all the features. Since these differences in variance have nothing to do with importance in the biometric sense, it makes no sense to allow them to have greater weight in the assessment of biometric performance. We employed two types of features: (1) real eye-movement features and (2) synthetic features. We compare six variance normalization methods (histogram equalization, L1-normalization, median normalization, z-score normalization, min–max normalization, and L-infinite normalization) and one distance metric (Mahalanobis distance) in terms of how well they reduce the impact of the variance differences. The effectiveness of different techniques on real data depended on the strength of the inter-correlation of the features. For weakly correlated real features and synthetic features, histogram equalization was the best method followed by L1 normalization.

Book ChapterDOI
01 Jan 2019
TL;DR: This paper presents mathematical description of different distance metrics which can be acquired with different clustering algorithm and comparing their performance using the number of iterations used in computing the objective function, the misclassification of the datum in the cluster, and error between ideal cluster center location and observed center location.
Abstract: In the process of clustering, our attention is to find out basic procedures that measures the degree of association between the variables. Many clustering methods use distance measures to find similarity or dissimilarity between any pair of objects. The fuzzy c-means clustering algorithm is one of the most widely used clustering techniques which uses Euclidean distance metrics as a similarity measurement. The choice of distance metrics should differ with the data and how the measure of their comparison is done. The main objective of this paper is to present mathematical description of different distance metrics which can be acquired with different clustering algorithm and comparing their performance using the number of iterations used in computing the objective function, the misclassification of the datum in the cluster, and error between ideal cluster center location and observed center location.

Journal ArticleDOI
TL;DR: This paper proposes a transfer learning framework called Semi-Supervised Metric Transfer Learning with Relative Constraints (SSMTR), that uses distance metric learning with a set of relative distance constraints that capture the similarities and dissimilarities between the source and the target domains better.
Abstract: Transfer Learning is an effective method of dealing with real-world problems where the training and test data are drawn from different distributions. Transfer learning methods use a labeled source domain to boost the task in a target domain that may be unsupervised or semi-supervised. However, the previous transfer learning algorithms use Euclidean distance or Mahalanobis distance formula to represent the relationships between instances and to try and capture the geometry of the manifold. In many real-world scenarios, this is not enough and these functions fail to capture the intrinsic geometry of the manifold that the data exists in. In this paper, we propose a transfer learning framework called Semi-Supervised Metric Transfer Learning with Relative Constraints (SSMTR), that uses distance metric learning with a set of relative distance constraints that capture the similarities and dissimilarities between the source and the target domains better. In SSMTR, instance weights are learned for different domains which are then used to reduce the domain shift while a Relative Distance metric is learned in parallel. We have developed SSMTR for classification problems as well, and have conducted extensive experiments on several real-world datasets; particularly, the PIE Face, Office-Caltech, and USPS-MNIST datasets to verify the accuracy of our proposed algorithm when compared to the current transfer learning algorithms.

Journal ArticleDOI
TL;DR: The proposed alignment framework for multimodal HR images not only can align the different multimodale data more accurately than existing state-of-the-art domain adaptation methods, but also has a fast and simple procedure for large-scale data situation which is caused by HR imaging.
Abstract: High-resolution (HR) remote sensing (RS) imaging opens the door to very accurate geometrical analysis for objects. However, it is difficult to simultaneous use massive HR RS images in practical applications, because these HR images are often collected in different multimodal conditions (multisource, multiarea, multitemporal, multiresolution, and multiangular) and learning method trained for one situation is difficult to use for others. The key problem is how to simultaneously tackle three main problems: spectral drift, spatial deformation, and band inconsistency. To deal with these problems, we propose an unsupervised tensorized principal component alignment framework in this paper. In this framework, local spatial–spectral patch data are used as basic units in order to achieve simultaneously multidimensional alignment. This framework seeks a domain-invariant tensor feature space by learning multilinear mapping functions which align the source tensor subspace with the target tensor subspace on different dimensions. In addition, an approach based on the Mahalanobis distance for dimensionality estimation of tensor subspace is proposed to determine best sizes of the aligned tensor subspace for reducing computational complexity. HR images from GF-1, GF-2, DEIMOS-2, WorldView-2, and WorldView-3 satellites are used to evaluate the performance. The experimental results show the following two points: first, the proposed alignment framework for multimodal HR images not only can align the different multimodal data more accurately than existing state-of-the-art domain adaptation methods, but also has a fast and simple procedure for large-scale data situation which is caused by HR imaging. Second, the proposed tensor dimensionality estimation method is an efficient technology for seeking the intrinsic dimensions of high-order data.

Journal ArticleDOI
TL;DR: In this article, a multi-layer neural network architecture is proposed for few-shot image recognition of novel categories, where the transfer of knowledge is carried out at the feature extraction and the classification levels distributed across the two training stages.
Abstract: This paper proposes a multi-layer neural network structure for few-shot image recognition of novel categories. The proposed multi-layer neural network architecture encodes transferable knowledge extracted from a large annotated dataset of base categories. This architecture is then applied to novel categories containing only a few samples. The transfer of knowledge is carried out at the feature-extraction and the classification levels distributed across the two training stages. In the first-training stage, we introduce the relative feature to capture the structure of the data as well as obtain a low-dimensional discriminative space. Secondly, we account for the variable variance of different categories by using a network to predict the variance of each class. Classification is then performed by computing the Mahalanobis distance to the mean-class representation in contrast to previous approaches that used the Euclidean distance. In the second-training stage, a category-agnostic mapping is learned from the mean-sample representation to its corresponding class-prototype representation. This is because the mean-sample representation may not accurately represent the novel category prototype. Finally, we evaluate the proposed network structure on four standard few-shot image recognition datasets, where our proposed few-shot learning system produces competitive performance compared to previous work. We also extensively studied and analyzed the contribution of each component of our proposed framework.

Proceedings ArticleDOI
18 Jul 2019
TL;DR: Zhang et al. as mentioned in this paper proposed three deep learning approaches that utilize Mahalanobis distance to model complex interactions among users, playlists, and songs using only their interaction data.
Abstract: In this paper, we aim to solve the automatic playlist continuation (APC) problem by modeling complex interactions among users, playlists, and songs using only their interaction data. Prior methods mainly rely on dot product to account for similarities, which is not ideal as dot product is not metric learning, so it does not convey the important inequality property. Based on this observation, we propose three novel deep learning approaches that utilize Mahalanobis distance. Our first approach uses user-playlist-song interactions, and combines Mahalanobis distance scores between (i) a target user and a target song, and (ii) between a target playlist and the target song to account for both the user's preference and the playlist's theme. Our second approach measures song-song similarities by considering Mahalanobis distance scores between the target song and each member song (i.e., existing song) in the target playlist. The contribution of each distance score is measured by our proposed memory metric-based attention mechanism. In the third approach, we fuse the two previous models into a unified model to further enhance their performance. In addition, we adopt and customize Adversarial Personalized Ranking (APR) for our three approaches to further improve their robustness and predictive capabilities. Through extensive experiments, we show that our proposed models outperform eight state-of-the-art models in two large-scale real-world datasets.

Journal ArticleDOI
TL;DR: The experiment results show that the convergence rate, accuracy and generalization ability of the proposed method are improved compared with the traditional RBF neural network with TS fuzzy model in Qiao et al. (2014) and the GA-BP (Genetic Algorithm-Back Propagation) model in Wang et-al (2016).

Journal ArticleDOI
TL;DR: This paper reviews the literature related to developing and improving MTS theory, and presents and analyzes the research results in terms of MD, SNR, Mahalanobis Space (MS), feature selection, threshold, multi-class MTS, and comparison with other methods.

Journal ArticleDOI
TL;DR: The model has the advantage that it conveniently quantifies the qualitative indices, and it can integrate the data source information to improve the multi-objective performance indices, so that it is very useful to apply multi-source data and prior knowledge to multi- objective optimization of the automatic train operation control system.
Abstract: The automatic train operation which integrates knowledge-based intelligent algorithm to develop safe and efficient control system has become one of the most important developing directions in the field of railway transit equipment. Multi-objective optimization is a strictly incompatible problem, and such contradiction is one of the main reasons that lead to the best multi-objective optimization difficult to achieve. In this paper, the multi-objective optimization feature information is transformed into the association function first, and then the matter-element theory is introduced to establish models for the speed trajectory to achieve the multi-objective optimization to fuse knowledge-based safety requirement constrained condition. Performance indices weighting of different performance in different stages are determined with the Hierarchical Mahalanobis distance method, and the decision speeds are calculated with goodness evaluation method. Taking Shanghai Railway Transit Equipment in China as a case study, this paper selected the multi-objective performance indices including passenger comfort, running stability, energy efficiency, and parking accuracy as objectives to support the decision-making. The multi-objective performance indices are evaluated by a field investigation and simulation. The test result shows that the comfort level, running stability, energy saving property, and parking accuracy are better than those derived by the traditional control algorithm. It indicates that the model has the advantage that it conveniently quantifies the qualitative indices, and it can integrate the data source information to improve the multi-objective performance indices, so that it is very useful to apply multi-source data and prior knowledge to multi-objective optimization of the automatic train operation control system.

Journal ArticleDOI
TL;DR: A novel multivariate signal denoising method that computes Mahalanobis distance measure at multiple data scales obtained from multivariate empirical mode decomposition (MEMD) algorithm is presented and it is proved that the proposed method is able to incorporate inherent correlation between multiple data channels in the Denoising process.
Abstract: A novel multivariate signal denoising method is presented that computes Mahalanobis distance measure at multiple data scales obtained from multivariate empirical mode decomposition (MEMD) algorithm. That enables joint multichannel data denoising directly in multidimensional space $\mathcal {R}^N$ where input signal resides, by employing interval thresholding on multiple data scales in $\mathcal {R}^N$ . We provide theoretical justification of using Mahalanobis distance at multiple scales obtained from MEMD and prove that the proposed method is able to incorporate inherent correlation between multiple data channels in the denoising process. The performance of the proposed method is verified on a range of synthetic and real world signals.

Journal ArticleDOI
TL;DR: The proposed procedure is based on a k-means algorithm in which the distance between the curves is measured with a metric that generalizes the Mahalanobis distance in Hilbert spaces, considering the correlation and the variability along all the components of the functional data.
Abstract: This paper proposes a clustering procedure for samples of multivariate functions in $$(L^2(I))^{J}$$ , with $$J\ge 1$$ . This method is based on a k-means algorithm in which the distance between the curves is measured with a metric that generalizes the Mahalanobis distance in Hilbert spaces, considering the correlation and the variability along all the components of the functional data. The proposed procedure has been studied in simulation and compared with the k-means based on other distances typically adopted for clustering multivariate functional data. In these simulations, it is shown that the k-means algorithm with the generalized Mahalanobis distance provides the best clustering performances, both in terms of mean and standard deviation of the number of misclassified curves. Finally, the proposed method has been applied to two case studies, concerning ECG signals and growth curves, where the results obtained in simulation are confirmed and strengthened.

Journal ArticleDOI
TL;DR: This work demonstrated the special ability of FT-NIR spectroscopy in combination with DA to examine inkjet-printed documents in a fast and non-destructive fashion.

Journal ArticleDOI
TL;DR: T2EQ is a practically feasible equivalence procedure based on the Mahalanobis distance with an internal equivalence margin for comparing dissolution profiles that meets the needs of both authorities and industry and exceeds the power of methods recently discussed in the literature.
Abstract: For some postapproval changes, the manufacturer has to demonstrate that the dissolution profile of the drug product before the change is statistically equivalent to the dissolution profile after the change. Guidelines suggest the so-called similarity factor f2 as standard approach for the equivalence analysis. f2 is a statistically questionable transformation of the Euclidean distance between both profile means and does not allow a control of the type I error rate. An alternative multivariate distance measure for quantifying the dissimilarity between both profile groups is the Mahalanobis distance. Current equivalence procedures based on the Mahalanobis distance implicate some practical problems in the dissolution context: either one chooses an exact method but the determination of a product independent equivalence margin will not be practically feasible or one chooses an approximate alternative that suffers from the bias of the Mahalanobis distance point estimate. This paper suggests the T2EQ approach for dissolution profile comparisons. T2EQ is a practically feasible equivalence procedure based on the Mahalanobis distance with an internal equivalence margin for comparing dissolution profiles. The equivalence margin is compliant with current dissolution guidelines. The operating characteristics (size, robustness, and power) are investigated via simulation: T2EQ meets the needs of both authorities and industry: not affected by the bias of the point estimate the type I error rate can be reliably controlled for various distribution assumptions and the power of T2EQ exceeds the power of methods recently discussed in the literature. These results were presented for the first time at CEN-ISBS 2017.

Journal ArticleDOI
TL;DR: The coarse-to-fine diagnosing strategy is proposed to determine the initial fault of rolling bearing and has the capability of estimating initial fault and determining degradation stages of bearing.
Abstract: The initial fault signal of rolling element bearing is extremely weak and could be easily masked by strong background noise. Different features of vibration signal can be different sensitivity to initial fault and performance degradation. Moreover, individual features cannot reflect bearing fault rationally and these features reveal non-monotonic behavior when the bearing condition deteriorates. A Health Indicator (HI) is proposed based on Mahalanobis Distance and Cumulative Sum (MD-CUMSUM). The time-frequency domain features extracted through Singular Value Decomposition based on Variational Mode Decomposition (VMD-SVD) and several optimal time domain features are used to calculate Mahalanobis Distances (MDs). The coarse-to-fine diagnosing strategy is proposed to determine the initial fault of rolling bearing. The obtained HI is utilized to estimate the different performance degradation stages of the bearing depending on the thresholds. This method is verified by utilizing two different experiments. The results demonstrate that the approach has the capability of estimating initial fault and determining degradation stages of bearing.

Journal ArticleDOI
TL;DR: In microgrids, where the holistically centralized techniques do not meet the speed requirement, regionalization can be a viable solution to improve and speed up protection and control applications.
Abstract: In microgrids, where the holistically centralized techniques do not meet the speed requirement, regionalization can be a viable solution to improve and speed up protection and control applications. This paper presents a fast regionalized approach to mitigate the microgrid voltage and frequency deviations simultaneously. The regionalization technique is on the basis of a multidimensional distance criterion known as Mahalanobis index which can be executed at several load levels in a parallel manner. In contrast to the existing district-based methods which lead to unified areas from the geographic point of view, the new technique regionalizes the network without enforcing geographic distance condition, with more degree of freedom. This feature assures faster and more effective voltage and frequency recovery. Since the load model inlfuences the voltage and frequency of the grid particularly in abnormal conditions, a hybrid load model dedicated to low-voltage distribution systems is employed here. To appraise and compare the performance and accuracy of the proposed plan with an existing one, the standard IEEE 37-bus distribution system is assayed. The steady state results and dynamic responses attest to the superiority of the proposed plan and guarantee satisfactory fulfillment for real-world practices.

Journal ArticleDOI
12 Oct 2019
TL;DR: In this paper, the authors provide an extensive study of two relevant manifold learning algorithms, empirical intrinsic geometry (EIG) and locally linear embedding (LLE) under the manifold setup.
Abstract: Local covariance structure under the manifold setup has been widely applied in the machine-learning community. Based on the established theoretical results, we provide an extensive study of two relevant manifold learning algorithms, empirical intrinsic geometry (EIG) and locally linear embedding (LLE) under the manifold setup. Particularly, we show that without an accurate dimension estimation, the geodesic distance estimation by EIG might be corrupted. Furthermore, we show that by taking the local covariance matrix into account, we can more accurately estimate the local geodesic distance. When understanding LLE based on the local covariance structure, its intimate relationship with the curvature suggests a variation of LLE depending on the “truncation scheme”. We provide a theoretical analysis of the variation.

Journal ArticleDOI
TL;DR: The Kernel Cross-View Collaborative Representation based Classification (Kernel X-CRC) is proposed that represents probe and gallery images by balancing representativeness and similarity nonlinearly and achieves state-of-the-art for rank-1 matching rates in two person re-identification datasets and the second best results on VIPeR and CUHK01 datasets.