Multiview Kernels for Low-Dimensional Modeling of Seismic Events

Lindenbaum, Ofir; Bregman, Yuri; Rabin, Neta; Averbuch, Amir

doi:10.1109/tgrs.2018.2797537

Cited by 20 publications

(14 citation statements)

References 51 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Moreover, thus far, there has been no relevant literature that reports on TEC precursors using K2DPCA. Reference [29] proposed a kernel-fusion based dimensionality combined reduction framework using the concept of kernel called Kernel Product (KP), Kernel Sum (KS) and Kernel Canonical Correlation Analysis (KCCA) to generate meaningful seismic representations from a non-linear one-dimensional seismic data set based on 2023 events in Israel and Jordan. PCA was used as a tool to reduce the dimension of the data while preserving most of the variance.…”

Section: B Related Studies Of Tec Precursorsmentioning

confidence: 99%

Detecting All Possible Ionospheric Precursors by Kernel-Based Two-Dimensional Principal Component Analysis

Lin

Chao

2019

IEEE Access

View full text Add to dashboard Cite

Kernel-based two-dimensional principal component analysis (K2DPCA), a nonlinear method, was performed to examine ionospheric 2-D total electron content (TEC) variations obtained from the NASA Global Differential GPS (GDGPS) network, which consisted of low spatial resolution ionospheric TEC data, to detect TEC precursors prior to the China Ludian earthquake at 08:30:13 UT on 03 August 2014 (M = 6.1). Simultaneously, two-dimensional principal component analysis (2DPCA), a linear method, was performed to examine data related to the same earthquake for comparison purposes. The results have shown that two TEC precursors were observed three days prior to this earthquake. The characteristics of all principal eigenvalues were considered to determine the TEC precursors of the earthquake in this paper. This paper aimed to differentiate the previous studies of Lin about 2DPCA using only the characteristic of the first principal eigenvalue to determine the TEC precursors of the earthquake. However, some other TEC precursors could be lost in processes that do not analyze the characteristics of other principal eigenvalues. Therefore, this research can be considered as a full analysis that examined the detailed surrounding TEC precursors of the earthquake and could be treated as a new study scope with 2DPCA and K2DPCA. The TEC precursor during 06:15 to 06:20 (UT) on 01 August 2014 localized near the epicenter was more intense compared with the characteristic of the first largest principal eigenvalue. However, another localized TEC precursor, which occurred from 06:15 to 06:20 (UT) on 01 August 2014 at a site west of the epicenter, was found to be more intense. The more plausible reason for the two TEC precursors could be the density variances caused by radon-gas release due to the fine crack stress transfer from the rocks according to Coulomb stress transfer theory as applied to a far TEC precursor. The durations of both TEC precursors were at least 5 min. This research could be treated as a full and detailed examination of surrounding TEC precursors. K2DPCA was better at identifying the TEC precursors. INDEX TERMS Kernel-based two-dimensional principal component analysis (K2DPCA), 2D total electron content (TEC), TEC precursors, NASA global differential GPS (GDGPS) network, two-dimensional principal component analysis (2DPCA), China Ludian earthquake, Radon-gas release, Coulomb stress transfer theory.

show abstract

Section: B Related Studies Of Tec Precursorsmentioning

confidence: 99%

Detecting All Possible Ionospheric Precursors by Kernel-Based Two-Dimensional Principal Component Analysis

Lin

Chao

2019

IEEE Access

View full text Add to dashboard Cite

show abstract

“…However, for the second sensor, due to the strong noise, A 2 will be close to a perturbed Gram matrix that mainly comes from the high dimensional noise. Consequently, as illustrated in (33), the NCCA matrix will be close to N which is a product matrix of the clean transition matrix and the shifted Gram matrix. Clearly, the clean transition matrix is contaminated by the shifted Gram matrix, which does not contain any information about the signal.…”

Section: Main Results (I)-classic Bandwidthmentioning

confidence: 99%

“…The main idea beyond these developments is that the nonlinear structure is modeled by various nonlinear geometric structures, and the algorithms are designed to preserve and capture this nonlinear structure. Such ideas and algorithms have been successfully applied to many real world problems, like audio-visual voice activity detection [10], the study of the sequential audio-visual correspondence [9], automatic sleep stage annotation from two electroencephalogram signals [35], seismic event modeling [33], fetal electrocardiogram analysis [42] and IQ prediction from two fMRI paradigms [48], which is a far from complete list.…”

Section: Introductionmentioning

confidence: 99%

On the Spectral Property of Kernel-Based Sensor Fusion Algorithms of High Dimensional Data

Ding

2021

IEEE Trans. Inform. Theory

View full text Add to dashboard Cite

We study the behavior of two kernel based sensor fusion algorithms, nonparametric canonical correlation analysis (NCCA) and alternating diffusion (AD), under the nonnull setting that the clean datasets collected from two sensors are modeled by a common low dimensional manifold embedded in a high dimensional Euclidean space and the datasets are corrupted by high dimensional noise. We establish the asymptotic limits and convergence rates for the eigenvalues of the associated kernel matrices assuming that the sample dimension and sample size are comparably large, where NCCA and AD are conducted using the Gaussian kernel. It turns out that both the asymptotic limits and convergence rates depend on the signal-to-noise ratio (SNR) of each sensor and selected bandwidths. On one hand, we show that if NCCA and AD are directly applied to the noisy point clouds without any sanity check, it may generate artificial information that misleads scientists' interpretation. On the other hand, we prove that if the bandwidths are selected adequately, both NCCA and AD can be made robust to high dimensional noise when the SNRs are relatively large.

show abstract

“…For instance, MLP was used to classify three classes of the Stromboli volcano, in Italy [37]. The LDA was tested for classifying seismic signals with the goal of differentiating earthquakes from man-made explosions [38]. The RF was used in the classification of earthquake and non-earthquake signals [39] and the SVM was used to perform classification of volcanic events [18].…”

Section: Classificationmentioning

confidence: 99%

Automatic Multichannel Volcano-Seismic Classification Using Machine Learning and EMD

Lara

Fernandes

Inza

et al. 2020

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

This article proposes the design of an automatic classifier using the empirical mode decomposition (EMD) along with machine learning techniques for identifying the five most important types of events of the Ubinas volcano, the most active volcano in Peru. The proposed method uses attributes from temporal, spectral and cepstral domains, extracted from the EMD of the signals, as well as a set of pre-processing and instrument correction techniques. Due to the fact that multichannel sensors are currently being installed in seismic networks worldwide, the proposed approach uses a multichannel sensor to perform the classification, contrary to the usual approach of the literature of using a single channel. The presented method is scalable to use data from multiple stations with one or more channels. The principal component analysis (PCA) method is applied to reduce the dimensionality of the feature vector and the supervised classification is carried out by means of several machine learning algorithms, the support vector machine (SVM) providing the best results. The presented investigation was tested with a large database that has a considerable number of explosion events, measured at the Ubinas volcano, located in Arequipa, Peru. The proposed classification system achieved a success rate of more than 90%.

show abstract

Multiview Kernels for Low-Dimensional Modeling of Seismic Events

Cited by 20 publications

References 51 publications

Detecting All Possible Ionospheric Precursors by Kernel-Based Two-Dimensional Principal Component Analysis

Detecting All Possible Ionospheric Precursors by Kernel-Based Two-Dimensional Principal Component Analysis

On the Spectral Property of Kernel-Based Sensor Fusion Algorithms of High Dimensional Data

Automatic Multichannel Volcano-Seismic Classification Using Machine Learning and EMD

Contact Info

Product

Resources

About