pith. sign in

arxiv: 2502.11212 · v2 · submitted 2025-02-16 · 📡 eess.SP

Non-negative tensor factorization-based dependence map analysis for local damage detection in presence of non-Gaussian noise

Anna Michalak , Justyna Hebda-Sobkowicz , Anil Kumar , Radoslaw Zimroz , Rafal Zdunek , Agnieszka Wylomanska This is my paper

Pith reviewed 2026-05-23 02:43 UTC · model grok-4.3

classification 📡 eess.SP
keywords non-negative tensor factorizationdependence mapinformative frequency banddamage detectionrolling element bearingsnon-Gaussian noisevibration analysistime-frequency map
0
0 comments X

The pith

Factorizing dependence maps with non-negative tensor factorization extracts informative frequency bands for bearing damage detection amid non-Gaussian noise.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops a technique to identify informative frequency bands in vibration signals from rolling element bearings by creating dependence maps that show spectral auto-similarity and then factorizing collections of these maps using non-negative tensor factorization. This factorization separates the components tied to damage from those caused by disturbances or other sources. The approach addresses difficulties from non-Gaussian noise, overlapping signatures, and low signal quality that hinder standard time-frequency analysis. If successful, it supplies a data-driven way to select the bands needed for reliable local damage detection without relying on manual tuning or assumptions about Gaussian noise.

Core claim

The method constructs dependence maps from time-frequency representations of the signal and applies non-negative tensor factorization to a tensor formed by multiple such maps. This process decomposes the data into informative and non-informative parts, with the informative parts corresponding to frequency bands that reveal local damage in the bearings.

What carries the argument

Non-negative tensor factorization of a collection of dependence maps, where each map encodes the auto-similarity structure of spectral content in the time-frequency plane.

If this is right

  • Informative frequency bands can be isolated even when non-Gaussian disturbances are present.
  • The extracted bands enable local damage detection in rolling element bearings.
  • The method handles overlapping fault signatures in the signals.
  • Validation occurs on both synthetic signals and real vibration data.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The technique might reduce the need for expert intervention in setting frequency band thresholds during monitoring.
  • It could extend to other types of machinery where vibration analysis is used for fault detection.
  • Combining this with other tensor methods might improve robustness under varying operating conditions.

Load-bearing premise

The auto-similarity patterns captured in dependence maps are distinct enough between damage-related content and non-informative noise that factorization can reliably separate them.

What would settle it

A test where the extracted components from NTF on dependence maps of a damaged bearing signal do not show higher energy or correlation at the known fault characteristic frequencies compared to a healthy signal.

Figures

Figures reproduced from arXiv: 2502.11212 by Agnieszka Wylomanska, Anil Kumar, Anna Michalak, Justyna Hebda-Sobkowicz, Radoslaw Zimroz, Rafal Zdunek.

Figure 1
Figure 1. Figure 1: Flowchart of the proposed method where γt−h is the shift window, xm = (x1, . . . , xP) is the m th segment of the input signal of length P, t = t1, . . . , tT is the time point, f = f1, . . . , fF is a frequency bin, and j is an imaginary unit; see [57] for more details. The spectrogram is defined as the absolute values of STFT, that is, S(f, t) = |S T FT(f, t)| ∈ R T×F . During data analysis, the commonly… view at source ↗
Figure 2
Figure 2. Figure 2: NTF factorization of the 3-order tensor C. In general, the NTF model for the three-order tensor C ∈ R F×F×M + can be defined as follows [54, 55]: C = X K k=1 (wk ◦ hk ◦ qk ) + E = I ×1 W ×2 H ×3 Q + E, (4) 8 [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Simulated signal 30 second length (a) with 1 second length signal fragments presented in (b) and (c), its [PITH_FULL_IMAGE:figures/full_fig_p015_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Pearson correlation maps for the simulated signal. [PITH_FULL_IMAGE:figures/full_fig_p016_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: NTF decomposition results: matrix H, N = 30, simulated signal: (a) feature vectors in H for all classes (b) Selectors based on feature vectors in H for all classes. The selector corresponding to the IFB is shown in a bold line. frequency band (2-3kHz). Even if the Pearson-based selector identifies this band (see [PITH_FULL_IMAGE:figures/full_fig_p016_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Filter characteristics: (a) NTF β = −1 (b) NTF β = 0 (c) NTF β = 1 for the simulated signal, filtration results (d) - (f), and SES of the filtered signals with given selectors (g) - (i), respectively non-cyclic impulses in the band around 6kHz. The filter illustrated in [PITH_FULL_IMAGE:figures/full_fig_p017_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Selectors results: (a) spectral kurtosis, (b) CV, (c) alpha, (d) Pearson for the simulated signal [PITH_FULL_IMAGE:figures/full_fig_p017_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Comparison of sorted ENVSI values of the simulated signal, filtered signals with the given selectors and [PITH_FULL_IMAGE:figures/full_fig_p018_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Copper ore crusher [69] In Case 2 the signal was obtained using a test rig. The setup included an electric motor, a gearbox, couplings, and two bearings, as presented in [PITH_FULL_IMAGE:figures/full_fig_p019_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Test rig used in the experiment. 4.1. Preliminary results for real signals The real signals were initially analyzed according to the same convention as the simulation signals. The raw signal and selected segments for Case 1 are shown in [PITH_FULL_IMAGE:figures/full_fig_p020_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Raw vibration signal of 30 second length corresponding to the Case 1 (a) with 1 second length signal [PITH_FULL_IMAGE:figures/full_fig_p021_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Raw vibration signal of 30 second length corresponding to the Case 2 (a) with 1 second length signal [PITH_FULL_IMAGE:figures/full_fig_p021_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Pearson correlation maps for the real vibration signal corresponding to Case 1. [PITH_FULL_IMAGE:figures/full_fig_p022_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: NTF decomposition results for Case 1: matrix [PITH_FULL_IMAGE:figures/full_fig_p023_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Selectors results for Case 1: (a) NTF β = −1 (b) NTF β = 0 (c) NTF β = 1 for the real vibration signal, filtration results (d) - (f), and SES of the filtered signals with given selectors (g) - (i), respectively. 4.2.2. Case 2: test rig The results obtained with the proposed approach are presented in [PITH_FULL_IMAGE:figures/full_fig_p023_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: NTF decomposition results for Case 2: matrix [PITH_FULL_IMAGE:figures/full_fig_p024_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: Selectors results for Case 2: (a) NTF β = −1 (b) NTF β = 0 (c) NTF β = 1 for the real vibration signal, filtration results (d) - (f), and SES of the filtered signals with given selectors (g) - (i), respectively. 24 [PITH_FULL_IMAGE:figures/full_fig_p024_17.png] view at source ↗
Figure 18
Figure 18. Figure 18: Selectors results for Case 1: (a) spectral kurtosis, (b) CV, (c) alpha (d) Pearson for the real vibration signal, [PITH_FULL_IMAGE:figures/full_fig_p025_18.png] view at source ↗
Figure 19
Figure 19. Figure 19: Selectors results for Case 2: (a) spectral kurtosis, (b) CV, (c) alpha (d) Pearson for the real vibration signal, [PITH_FULL_IMAGE:figures/full_fig_p026_19.png] view at source ↗
Figure 20
Figure 20. Figure 20: Comparison of sorted ENVSI values of the real signals signal, filtered signals with the given selectors and [PITH_FULL_IMAGE:figures/full_fig_p027_20.png] view at source ↗
Figure 21
Figure 21. Figure 21: Efficiency of the proposed approach for three considered β parameters and different values of ACI, ANCI based on 50 MC simulations. algorithm. It can also be seen that the computation time is the longest for β = −1 in each case. This difference is particularly noticeable with larger nfft values and longer signal time. In comparison, the values of the computation time for β = 0 and β = 1 are noticeably low… view at source ↗
Figure 22
Figure 22. Figure 22: Computation time depends on the nfft and the signal length (T) for different values of the parameter β. 6. Conclusions The approach outlined in this work provides a novel methodology for detecting local damage in rolling element bearings using vibration signals. The proposed method demonstrates adaptability and potential applicability in various real-world scenarios by addressing the challenges of weak cy… view at source ↗
Figure 23
Figure 23. Figure 23: The results (a) and SES of the filtered signals with the ENVSI value (b) of simulated signal analysis by the [PITH_FULL_IMAGE:figures/full_fig_p033_23.png] view at source ↗
Figure 24
Figure 24. Figure 24: The results (a) and SES of the filtered signals with the ENVSI value (b) of Case 1 analysis by the following [PITH_FULL_IMAGE:figures/full_fig_p034_24.png] view at source ↗
Figure 25
Figure 25. Figure 25: The results (a) and SES of the filtered signals with the ENVSI value (b) of Case 2 analysis by the following [PITH_FULL_IMAGE:figures/full_fig_p034_25.png] view at source ↗
Figure 26
Figure 26. Figure 26: Efficiency of the selectors: spectral kurtosis, CV, alpha, and Pearson for different values of ACI, ANCI based on 50 MC simulations. (a) (b) (c) [PITH_FULL_IMAGE:figures/full_fig_p035_26.png] view at source ↗
Figure 27
Figure 27. Figure 27: Efficiency of the kurtogram, infogram, and CFFs-gram for different values of ACI, ANCI based on 50 MC simulations. 35 [PITH_FULL_IMAGE:figures/full_fig_p035_27.png] view at source ↗
read the original abstract

The time-frequency map (TFM) is frequently used in condition monitoring, necessitating further processing to select an informative frequency band (IFB) or directly detect damage. However, selecting an IFB is challenging due to the complexity of spectral structures, non-Gaussian disturbances, and overlapping fault signatures in vibration signals. Additionally, dynamic operating conditions and low signal-to-noise ratio further complicate the identification of relevant features that indicate damage. To solve this problem, the present work proposes a novel method for informative band selection and local damage detection in rolling element bearings, utilizing non-negative tensor factorization (NTF)-based dependence map analysis. The recently introduced concept of the dependence map is leveraged, with a set of these maps being factorized to separate informative components from non-informative ones. Dependence maps provide valuable information on the auto-similarity of spectral content, while NTF, a powerful tool commonly used in image processing for feature extraction, enhances this process. The combination of these methods allows for the extraction of IFBs, forming the basis for local damage detection. The effectiveness of the proposed method has been validated using both synthetic and real vibration signals corrupted with non-Gaussian disturbances.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 0 minor

Summary. The paper proposes a method for informative frequency band (IFB) selection and local damage detection in rolling element bearings that applies non-negative tensor factorization (NTF) to a collection of dependence maps derived from time-frequency representations of vibration signals. The approach is presented as a way to separate informative from non-informative components even when non-Gaussian disturbances and overlapping fault signatures are present. Effectiveness is asserted on the basis of validation experiments performed on both synthetic and real vibration signals.

Significance. If the central claim holds, the combination of dependence-map construction with NTF factorization could supply a practical route to IFB extraction under realistic noise conditions that are difficult for conventional spectral methods. The manuscript does not, however, supply the quantitative evidence, baseline comparisons, or implementation details needed to evaluate whether that potential is realized.

major comments (1)
  1. [Abstract] Abstract: the claim that the method 'has been validated using both synthetic and real vibration signals' is presented without any accompanying quantitative results, error metrics, comparison baselines, or description of the experimental protocol, rendering the central effectiveness claim impossible to assess from the supplied text.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the comment on the abstract. We respond below.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the claim that the method 'has been validated using both synthetic and real vibration signals' is presented without any accompanying quantitative results, error metrics, comparison baselines, or description of the experimental protocol, rendering the central effectiveness claim impossible to assess from the supplied text.

    Authors: The abstract provides a high-level summary of the validation. Detailed quantitative results (including detection accuracy, false positive rates under non-Gaussian noise), error metrics, baseline comparisons (e.g., against spectral kurtosis and other IFB methods), and full experimental protocols appear in Sections 4 (synthetic signals) and 5 (real bearing data) of the manuscript. We will revise the abstract to incorporate key quantitative highlights and a brief protocol description. revision: yes

Circularity Check

0 steps flagged

No significant circularity

full rationale

The provided abstract and context describe a novel methodological proposal that combines dependence maps with non-negative tensor factorization (NTF) to extract informative frequency bands (IFBs) for damage detection. No equations, derivations, or load-bearing steps are shown that reduce any claimed result to a fitted parameter, self-definition, or self-citation chain. The central claim is presented as an empirical validation on synthetic and real signals rather than a mathematical reduction to inputs by construction. This matches the default expectation for papers without exhibited circularity.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the domain assumption that dependence maps encode useful auto-similarity information separable by NTF. No free parameters or invented entities are mentioned.

axioms (1)
  • domain assumption Dependence maps provide valuable information on the auto-similarity of spectral content
    Explicitly stated in the abstract as the foundation for applying NTF.

pith-pipeline@v0.9.0 · 5764 in / 1293 out tokens · 54772 ms · 2026-05-23T02:43:59.853260+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

69 extracted references · 69 canonical work pages

  1. [1]

    R. B. Randall, J. Antoni, Rolling element bearing diagnostics - a tutorial, Mechanical Systems and Signal Processing 25 (2) (2011) 485–520

    work page 2011

  2. [2]

    Antoni, R

    J. Antoni, R. Randall, The spectral kurtosis: application to the vibratory surveillance and diagnostics of rotating machines, Mechanical Systems and Signal Processing 20 (2) (2006) 308–331

    work page 2006

  3. [3]

    Antoni, Fast computation of the kurtogram for the detection of transient faults, Mechanical Systems and Signal Processing 21 (1) (2007) 108–124

    J. Antoni, Fast computation of the kurtogram for the detection of transient faults, Mechanical Systems and Signal Processing 21 (1) (2007) 108–124

    work page 2007

  4. [4]

    D. Wang, W. T. Peter, K. L. Tsui, An enhanced Kurtogram method for fault diagnosis of rolling element bearings, Mechanical Systems and Signal Processing 35 (1-2) (2013) 176–199

    work page 2013

  5. [5]

    Two automatic vibration-based fault diagnostic methods using the novel sparsity measurement–Parts 1 and 2

    W. T. Peter, D. Wang, The design of a new sparsogram for fast bearing fault diagnosis: Part 1 of the two related manuscripts that have a joint title as “Two automatic vibration-based fault diagnostic methods using the novel sparsity measurement–Parts 1 and 2”, Mechanical Systems and Signal Processing 40 (2) (2013) 499–519

    work page 2013

  6. [6]

    T. Wang, Q. Han, F. Chu, Z. Feng, A new skrgram based demodulation technique for planet bearing fault detection, Journal of Sound and Vibration 385 (2016) 330–349

    work page 2016

  7. [7]

    Z. Liu, Y . Jin, M. J. Zuo, D. Peng, Accugram: A novel approach based on classification to frequency band selection for rotating machinery fault diagnosis, ISA Transactions 95 (2019) 346–357

    work page 2019

  8. [8]

    Moshrefzadeh, A

    A. Moshrefzadeh, A. Fasana, The Autogram: An effective approach for selecting the optimal demodulation band in rolling element bearings diagnosis, Mechanical Systems and Signal Processing 105 (2018) 294–318

    work page 2018

  9. [9]

    Mauricio, W

    A. Mauricio, W. A. Smith, R. B. Randall, J. Antoni, K. Gryllias, Improved envelope spectrum via feature optimisation-gram (iesfogram): A novel tool for rolling element bearing diagnostics under non-stationary oper- ating conditions, Mechanical Systems and Signal Processing 144 (2020) 106891

    work page 2020

  10. [10]

    K. Wu, N. Chu, D. Wu, J. Antoni, The Enkurgram: A characteristic frequency extraction method for fluid machinery based on multi-band demodulation strategy, Mechanical Systems and Signal Processing 155 (2021) 107564

    work page 2021

  11. [11]

    G. Yu, C. Li, J. Zhang, A new statistical modeling and detection method for rolling element bearing faults based on alpha-stable distribution, Mechanical Systems and Signal Processing 41 (1-2) (2013) 155 – 175

    work page 2013

  12. [12]

    Y . Miao, M. Zhao, J. Lin, Improvement of kurtosis-guided-grams via gini index for bearing fault feature identi- fication, Measurement Science and Technology 28 (12) (2017) 125001

    work page 2017

  13. [13]

    Hebda-Sobkowicz, R

    J. Hebda-Sobkowicz, R. Zimroz, M. Pitera, A. Wyłoma ´nska, Informative frequency band selection in the pres- ence of non-gaussian noise – a novel approach based on the conditional variance statistic with application to bearing fault diagnosis, Mechanical Systems and Signal Processing 145 (2020) 106971

    work page 2020

  14. [14]

    Wodecki, A

    J. Wodecki, A. Michalak, R. Zimroz, Optimal filter design with progressive genetic algorithm for local damage detection in rolling bearings, Mechanical Systems and Signal Processing 102 (2018) 102–116

    work page 2018

  15. [15]

    Antoni, The infogram: Entropic evidence of the signature of repetitive transients, Mechanical Systems and Signal Processing 74 (2016) 73–94

    J. Antoni, The infogram: Entropic evidence of the signature of repetitive transients, Mechanical Systems and Signal Processing 74 (2016) 73–94

    work page 2016

  16. [16]

    Wang, An extension of the infograms to novel Bayesian inference for bearing fault feature identification, Mechanical Systems and Signal Processing 80 (2016) 19–30

    D. Wang, An extension of the infograms to novel Bayesian inference for bearing fault feature identification, Mechanical Systems and Signal Processing 80 (2016) 19–30

    work page 2016

  17. [17]

    Hebda-Sobkowicz, R

    J. Hebda-Sobkowicz, R. Zimroz, A. Wyłoma ´nska, J. Antoni, Infogram performance analysis and its enhance- ment for bearings diagnostics in presence of non-gaussian noise, Mechanical Systems and Signal Processing 170 (2022) 108764

    work page 2022

  18. [18]

    Cheng, S

    Y . Cheng, S. Wang, B. Chen, G. Mei, W. Zhang, H. Peng, G. Tian, An improved envelope spectrum via candi- date fault frequency optimization-gram for bearing fault diagnosis, Journal of Sound and Vibration 523 (2022) 116746

    work page 2022

  19. [19]

    Borghesani, J

    P. Borghesani, J. Antoni, CS2 analysis in presence of non-Gaussian background noise – E ffect on traditional estimators and resilience of log-envelope indicators, Mechanical Systems and Signal Processing 90 (2017) 378– 398

    work page 2017

  20. [20]

    Mauricio, J

    A. Mauricio, J. Qi, W. Smith, M. Sarazin, R. Randall, K. Janssens, K. Gryllias, Bearing diagnostics under strong electromagnetic interference based on Integrated Spectral Coherence, Mechanical Systems and Signal Processing 140 (2020) 106673

    work page 2020

  21. [21]

    Hebda-Sobkowicz, R

    J. Hebda-Sobkowicz, R. Zimroz, A. Wyłoma´nska, Selection of the informative frequency band in a bearing fault 30 diagnosis in the presence of non-Gaussian noise – comparison of recently developed methods, Applied Sciences 10 (8) (2020) 2657

    work page 2020

  22. [22]

    Kruczek, R

    P. Kruczek, R. Zimroz, A. Wyłoma´nska, How to detect the cyclostationarity in heavy-tailed distributed signals, Signal Processing 172 (2020) 107514

    work page 2020

  23. [23]

    N. Zhou, Y . Cheng, Z. Wang, B. Chen, W. Zhang, C ffsgram: A candidate fault frequencies-based optimal demodulation band selection method for axle-box bearing fault diagnosis, Measurement 207 (2023) 112368

    work page 2023

  24. [24]

    L. Cui, X. Zhao, D. Liu, H. Wang, A spectral coherence cyclic periodic index optimization-gram for bearing fault diagnosis, Measurement 224 (2024) 113898

    work page 2024

  25. [25]

    Y . Yang, R. Xie, M. Li, W. Cheng, A review on the application of blind source separation in vibration analysis of mechanical systems, Measurement (2024) 114241

    work page 2024

  26. [26]

    J. Allen, Short term spectral analysis, synthesis, and modification by discrete Fourier transform, IEEE Transac- tions on Acoustics, Speech, and Signal Processing 25 (3) (1977) 235–238

    work page 1977

  27. [27]

    Millioz, N

    F. Millioz, N. Martin, Circularity of the STFT and spectral kurtosis for time-frequency segmentation in gaussian environment, IEEE Trans. Signal Process. 59 (2) (2011) 515–524

    work page 2011

  28. [28]

    Huillery, F

    J. Huillery, F. Millioz, N. Martin, On the description of spectrogram probabilities with a chi-squared law, IEEE Transactions on Signal Processing 56 (6) (2008) 2249–2258

    work page 2008

  29. [29]

    Wodecki, A

    J. Wodecki, A. Michalak, R. Zimroz, T. Barszcz, A. Wyłoma ´nska, Impulsive source separation using combi- nation of nonnegative matrix factorization of bi-frequency map, spatial denoising and monte carlo simulation, Mechanical Systems and Signal Processing 127 (2019) 89–101

    work page 2019

  30. [30]

    Wodecki, A

    J. Wodecki, A. Michalak, R. Zimroz, A. Wyłoma´nska, Separation of multiple local-damage-related components from vibration data using nonnegative matrix factorization and multichannel data fusion, Mechanical Systems and Signal Processing 145 (2020) 106954

    work page 2020

  31. [31]

    Wodecki, A

    J. Wodecki, A. Michalak, R. Zimroz, Local damage detection based on vibration data analysis in the presence of gaussian and heavy-tailed impulsive noise, Measurement 169 (2021) 108400

    work page 2021

  32. [32]

    Gabor, R

    M. Gabor, R. Zdunek, R. Zimroz, A. Wylomanska, Bearing damage detection with orthogonal and nonnegative low-rank feature extraction, IEEE Transactions on Industrial Informatics 20 (2) (2024) 2944 – 2955

    work page 2024

  33. [33]

    D. D. Lee, H. S. Seung, Learning the parts of objects by non-negative matrix factorization, nature 401 (6755) (1999) 788–791

    work page 1999

  34. [34]

    C. Wang, M. Gan, et al., Non-negative emd manifold for feature extraction in machinery fault diagnosis, Mea- surement 70 (2015) 188–202

    work page 2015

  35. [35]

    Liang, F

    L. Liang, F. Liu, M. Li, K. He, G. Xu, Feature selection for machine fault diagnosis using clustering of non- negation matrix factorization, Measurement 94 (2016) 295–305

    work page 2016

  36. [36]

    X. Wang, Y . Cai, A. Li, W. Zhang, Y . Yue, A. Ming, Intelligent fault diagnosis of diesel engine via adaptive vmd-rihaczek distribution and graph regularized bi-directional nmf, Measurement 172 (2021) 108823

    work page 2021

  37. [37]

    Shashua, T

    A. Shashua, T. Hazan, Non-negative tensor factorization with applications to statistics and computer vision, in: Proc. of the 22-th International Conference on Machine Learning, Bonn, Germany, 2005

    work page 2005

  38. [38]

    J. D. Carroll, G. De Soete, S. Pruzansky, Fitting of the latent class model via iteratively reweighted least squares candecomp with nonnegativity constraints, in: Multiway data analysis, 1989, pp. 463–472

    work page 1989

  39. [39]

    FitzGerald, M

    D. FitzGerald, M. Cranitch, E. Coyle, Extended nonnegative tensor factorisation models for musical sound source separation, Comput Intell Neurosci 872425 (2008) 1–15

    work page 2008

  40. [40]

    Figueiredo, B

    M. Figueiredo, B. Ribeiro, A. de Almeida, Electrical signal source separation via nonnegative tensor factor- ization using on site measurements in a smart home, IEEE Transactions on Instrumentation and Measurement 63 (2) (2014) 364–373

    work page 2014

  41. [41]

    Xiong, Y

    F. Xiong, Y . Qian, J. Zhou, Y . Y . Tang, Hyperspectral unmixing via total variation regularized nonnegative tensor factorization, IEEE Transactions on Geoscience and Remote Sensing 57 (4) (2019) 2341–2357

    work page 2019

  42. [42]

    Liang, H

    L. Liang, H. Wen, F. Liu, G. Li, M. Li, Feature extraction of impulse faults for vibration signals based on sparse non-negative tensor factorization, Applied Sciences 9 (18) (2019)

    work page 2019

  43. [43]

    Gabor, R

    M. Gabor, R. Zdunek, R. Zimroz, J. Wodecki, A. Wylomanska, Non-negative tensor factorization for vibration- based local damage detection, Mechanical Systems and Signal Processing 198 (2023)

    work page 2023

  44. [44]

    D. Zhao, D. Shao, L. Cui, Ctnet: A data-driven time-frequency technique for wind turbines fault diagnosis under time-varying speeds, ISA Transactions 154 (2024) 335 – 351. 31

    work page 2024

  45. [45]

    D. Zhao, H. Wang, L. Cui, Frequency-chirprate synchrosqueezing-based scaling chirplet transform for wind turbine nonstationary fault feature time–frequency representation, Mechanical Systems and Signal Processing 209 (2024) 111112

    work page 2024

  46. [46]

    Nowicki, J

    J. Nowicki, J. Hebda-Sobkowicz, R. Zimroz, A. Wyłoma ´nska, Dependency measures for the diagnosis of local faults in application to the heavy-tailed vibration signal, Applied Acoustics 178 (2021) 107974

    work page 2021

  47. [47]

    Wodecki, A

    J. Wodecki, A. Michalak, A. Wyłoma ´nska, R. Zimroz, Influence of non-gaussian noise on the e ffectiveness of cyclostationary analysis–simulations and real data analysis, Measurement 171 (2021) 108814

    work page 2021

  48. [48]

    Zheng, Z

    X. Zheng, Z. He, J. Nie, P. Li, Z. Dong, M. Gao, A progressive multi-source domain adaptation method for bearing fault diagnosis, Applied Acoustics 216 (2024) 109797

    work page 2024

  49. [49]

    M. Ye, X. Yan, N. Chen, M. Jia, Intelligent fault diagnosis of rolling bearing using variational mode extraction and improved one-dimensional convolutional neural network, Applied Acoustics 202 (2023) 109143

    work page 2023

  50. [50]

    Wißbrock, Z

    P. Wißbrock, Z. Ren, D. Pelkmann, More than spectrograms: Deep representation learning for machinery fault detection, Applied Acoustics 225 (2024) 110178

    work page 2024

  51. [51]

    T. Han, L. Zhang, Z. Yin, A. C. Tan, Rolling bearing fault diagnosis with combined convolutional neural net- works and support vector machine, Measurement: Journal of the International Measurement Confederation 177 (2021) 109022

    work page 2021

  52. [52]

    X. Li, H. Shao, S. Lu, J. Xiang, B. Cai, Highly e fficient fault diagnosis of rotating machinery under time- varying speeds using LSISMM and small infrared thermal images, IEEE Transactions on Systems, Man, and Cybernetics: Systems 52 (12) (2022) 7328 – 7340

    work page 2022

  53. [53]

    J. Luo, H. Shao, J. Lin, B. Liu, Meta-learning with elastic prototypical network for fault transfer diagnosis of bearings under unstable speeds, Reliability Engineering and System Safety 245 (2024) 110001

    work page 2024

  54. [54]

    Cichocki, R

    A. Cichocki, R. Zdunek, A. H. Phan, S.-i. Amari, Nonnegative matrix and tensor factorizations: applications to exploratory multi-way data analysis and blind source separation, John Wiley & Sons, 2009

    work page 2009

  55. [55]

    T. G. Kolda, B. W. Bader, Tensor decompositions and applications, SIAM Review 51 (3) (2009) 455–500

    work page 2009

  56. [56]

    Z. Feng, M. Liang, F. Chu, Recent advances in time–frequency analysis methods for machinery fault diagnosis: a review with application examples, Mechanical Systems and Signal Processing 38 (1) (2013) 165–205

    work page 2013

  57. [57]

    Boashash, Time-frequency signal analysis and processing: a comprehensive reference, Academic press, 2015

    B. Boashash, Time-frequency signal analysis and processing: a comprehensive reference, Academic press, 2015

    work page 2015

  58. [58]

    O. J. Dunn, V . A. Clark, Basic statistics: a primer for the biomedical sciences, John Wiley & Sons, 2009

    work page 2009

  59. [59]

    M. W. Berry, M. Browne, A. N. Langville, V . P. Pauca, R. J. Plemmons, Algorithms and applications for ap- proximate nonnegative matrix factorization, Computational statistics & data analysis 52 (1) (2007) 155–173

    work page 2007

  60. [60]

    S. Sra, I. Dhillon, Generalized nonnegative matrix approximations with bregman divergences, Advances in neural information processing systems 18 (2005)

    work page 2005

  61. [61]

    P. H. Westfall, Kurtosis as peakedness, 1905–2014. r.i.p., The American Statistician 68 (3) (2014) 191–195

    work page 1905

  62. [62]

    Joanes, C

    D. Joanes, C. Gill, Comparing measures of sample skewness and kurtosis, Journal of the Royal Statistical Soci- ety: Series D (The Statistician) 47 (1) (1998) 183–189

    work page 1998

  63. [63]

    Pachaud, R

    C. Pachaud, R. Salvetat, C. Fray, Crest factor and kurtosis contributions to identify defects inducing periodical impulsive forces, Mechanical Systems and Signal Processing 11 (6) (1997) 903–916

    work page 1997

  64. [64]

    Barszcz, A

    T. Barszcz, A. Jabło ´nski, A novel method for the optimal band selection for vibration signal demodulation and comparison with the kurtogram, Mechanical Systems and Signal Processing 25 (1) (2011) 431 – 451

    work page 2011

  65. [65]

    Samoradnitsky, Stable Non-Gaussian Random Processes: Stochastic Models with Infinite Variance, Chapman & Hall, 1994

    G. Samoradnitsky, Stable Non-Gaussian Random Processes: Stochastic Models with Infinite Variance, Chapman & Hall, 1994

    work page 1994

  66. [66]

    J. H. McCulloch, Simple consistent estimators of stable distribution parameters, Communications in Statistics- Simulation and Computation 15 (4) (1986) 1109–1136

    work page 1986

  67. [67]

    Jaworski, M

    P. Jaworski, M. Pitera, The 20-60-20 rule., Discrete & Continuous Dynamical Systems-Series B 21 (4) (2016)

    work page 2016

  68. [68]

    Jelito, M

    D. Jelito, M. Pitera, New fat-tail normality test based on conditional second moments with applications to finance, Available at SSRN 3303132 (2018)

    work page 2018

  69. [69]

    Wyłoma ´nska, R

    A. Wyłoma ´nska, R. Zimroz, J. Janczura, J. Obuchowski, Impulsive noise cancellation method for copper ore crusher vibration signals enhancement, IEEE Transactions on Industrial Electronics 63 (9) (2016) 5612–5621. 32 Appendix The comparative results popular and newly proposed methods for all analyzed signals are presented below. For the simulated data th...

    work page 2016