Tunnelling Through Time Series: A Probabilistic Visibility Graph for Local and Global Pattern Discovery

Jose Sanchez-Bornot; Roberto Sotero

REVIEW 5 major objections 5 minor 22 references

Tunnelling Through Time Series: A Probabilistic Visibility Graph for Local and Global Pattern Discovery

T0 review · 5 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The Probabilistic Visibility Graph adds tunnelling-style probabilistic links to the classical visibility graph, letting it capture long-range time-series patterns and distinguish rest from anesthesia in brain recordings.

desk verdict A clearly-defined probabilistic extension of the visibility graph, but the rest/anesthesia claim needs parameter reporting and sensitivity analysis before it convinces. read the letter →

arxiv 2507.01247 v1 pith:U5RQMMLV submitted 2025-07-01 eess.SY cs.SYphysics.data-an

Roberto Sotero , Jose Sanchez-Bornot This is my paper · ORCID

classification eess.SYcs.SYphysics.data-an

keywords Time-SeriesComplexNetworksVisibilityGraphProbabilisticlong-rangedependenciessmall-worldnetworkscale-freeelectrocorticography

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper introduces the Probabilistic Visibility Graph (PVG), a way of turning a time series into a network in which every pair of time points is connected with a probability that decays with the height of the intermediate values blocking the sight line, inspired by quantum tunnelling. The intent is to overcome the classical Visibility Graph's tendency to lose long-range relationships: with the PVG, distant points can still be linked, with strength tunable by a decay parameter and a threshold. The paper shows on simulated amplitude-modulated signals that the PVG reaches far larger maximum degrees than the classical VG, and on macaque electrocorticography data that the PVG's network metrics distinguish rest from anesthesia, with rest showing stronger small-world and scale-free (hub-dominated) topology. If the method holds up, it offers a flexible graph-based tool for capturing cross-scale, long-range patterns in long physiological and other time series.

What carries the argument

The load-bearing object is the probabilistic connection rule $P_{ij}=e^{-\rho h_{\max}}$ combined with the arctangent interaction strength $W_{ij}$ and the threshold $P_0$. The maximum obstruction height $h_{\max}$ is the tallest vertical gap between the straight line joining two time points and the intermediate data points; when no obstruction exists, $h_{\max}=0$ and $P_{ij}=1$. This rule is what lets the graph 'tunnel' through barriers that would block a classical visibility edge, with $\rho$ controlling how strongly obstructions suppress long-range links and $P_0$ controlling how many weak links survive to appear in the final graph. The same machinery, varied through $\rho$ and $P_0$, generates the full family of networks used in the simulations and the ECoG analysis.

What would settle it

Re-run the ECoG rest-versus-anesthesia analysis with a different decay kernel, such as $P_{ij}=h_{\max}^{-\alpha}$, while matching the network density between conditions; if the small-worldness and scale-freeness separation between rest and anesthesia disappears or reverses, the claim would be shown to depend on the chosen tunnelling model rather than on the structure of the brain signals.

Watch

Extended reading notes

Core claim

The central claim is that replacing the all-or-nothing visibility rule by a probabilistic one yields a network representation that preserves the classical local structure while adding long-range edges that carry global pattern information. Formally, after normalizing the series to $[0,1]$, the paper defines the connection probability $P_{ij}=e^{-\rho h_{\max}}$, where $h_{\max}$ is the largest obstruction height between points $i$ and $j$, and merges this with the weighted-visibility interaction $W_{ij}=\arctan((x(t_j)-x(t_i))/(t_j-t_i))$ to obtain the weighted connectivity $M_{ij}=W_{ij}P_{ij}$. A threshold $P_0$ converts $M$ into an adjacency matrix; $P_0=1$ recovers the classical VG and $P_0=0$ connects all pairs. The paper reports that on an amplitude-modulated test signal the PVG's maximum degree climbs to the theoretical maximum of about 5000, versus 56 for the classical VG, and that on ECoG data from macaques, rest exhibits higher clustering, shorter path lengths, stronger small-worldness, and scale-free degree distributions relative to anesthesia, which appears denser, more homogeneous, and less hub-dominated.

Load-bearing premise

The load-bearing premise is that the chance of a non-visible connection should drop off exponentially with the tallest obstruction height between the two time points, and that the user-chosen decay $\rho$ and threshold $P_0$ are the right way to balance local and global connections; if a different decay rule were used, the network topology and the reported rest versus anesthesia differences could change.

Editorial extensions

If this is right

Varying $P_0$ between 0 and 1 sweeps a family of graphs from the all-to-all network to the classical VG, so the PVG offers a tunable middle ground between local and global connectivity.
The simulated AM signal shows that the PVG captures long-range interactions that the classical VG misses, indicated by the maximum degree rising to the theoretical maximum (about 5000) instead of 56.
On real ECoG data, rest and anesthesia produce different PVG topologies: rest is small-world and scale-free (hub-dominated, short paths, high clustering) while anesthesia is denser and more homogeneous.
Network measures of the PVG vary systematically with $\rho$ and $P_0$, so conclusions about brain-state differences must be made over a range of parameters rather than at a single setting.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A natural extension the paper does not pursue is to test whether the rest/anesthesia separation is tied to the exponential decay form; re-running the analysis with a power-law or logistic decay kernel would show whether the reported differences are a property of the signals or of the chosen tunnelling model.
Because the AM simulation is explicitly designed to mimic phase-amplitude coupling, the PVG could be used as a network-based index of cross-frequency coupling strength, complementing spectral PAC measures.
The reported scale-free behaviour rests on a linear fit to the log-transformed degree distribution; applying stricter goodness-of-fit criteria would likely sharpen (or weaken) the claim that rest is genuinely scale-free.
The method's $O(N^2)$ cost limits direct application to very long series; approximate or hierarchical versions of the visibility check would be needed for terabyte-scale recordings.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Two things to know. First, the paper does deliver a new construction: replacing the hard visibility rule with the probabilistic edge probability P_ij = exp(-rho h_max) makes the visibility graph tunable between the classical VG (P0=1) and a fully connected graph, and the method is clearly written. Second, the empirical payoff is still conditional: the rest/anesthesia separation in ECoG is the headline result, but the threshold P0 used in that analysis is never reported, and there is no sensitivity analysis showing the ordering of the conditions survives changes in P0 or in the stipulated exponential decay.

The approach is genuinely simple and the simulations behave as one would hope: increasing rho lengthens paths, lowers clustering, and shrinks the maximum degree, while smaller P0 admits more long-range edges. Applying the PVG to real ECoG data with rest and anesthesia is a reasonable first test, and the idea that anesthesia homogenizes the network is plausible.

The soft spots are real but not fatal. The main one: Figure 3 plots everything as a function of rho, but the reader cannot know which P0 produced those curves, nor whether other P0 values flip the rest/anesthesia order. Since Eq. (4) is a postulate, a different decay function could reorder the results. Second, "rest showing significantly higher sigma" is asserted without any statistical test; the shaded standard deviations are informative but not a test. Third, the power-law exponent is from a linear fit to log-binned degree distributions; that is a crude estimator, and the paper cites Broido and Clauset, who show that most such claims do not survive rigorous testing. Fourth, no code or data are released, so the analysis is not reproducible as it stands.

If the authors report the exact parameters, add a sensitivity sweep over P0 and perhaps the decay family, and run a simple permutation test on the ECoG contrast, the paper would be solid as a methods note. As is, it is a plausible increment, not an established one. I would send it to peer review because the method is well-defined and the application is of interest, but I would expect major revision. Good for a signal-processing audience who wants a tunable VG; skip if you need a validated biomarker. My own take: I wouldn't cite the empirical claim, but I'd keep the construction in mind.

Referee Report

5 major / 5 minor

Summary. The paper introduces the Probabilistic Visibility Graph (PVG), an extension of the classical Visibility Graph in which nodes are connected with probability P_ij = exp(-rho h_max), where h_max is the maximum obstruction height and rho is a decay parameter; a threshold P0 converts these probabilities into a binary adjacency matrix. The authors test the method on a simulated amplitude-modulated signal, showing how network measures L, C, and k_max vary with rho and P0, and on macaque ECoG recordings under rest and anesthesia, reporting that rest exhibits higher clustering, stronger small-worldness, and more scale-free degree distributions than anesthesia. The central claim is that the PVG captures both local and global temporal patterns better than the classical VG, thereby offering a tool for analyzing multiscale time series.

Significance. If the empirical claims are supported, the PVG would be a useful, easily implementable extension of the visibility-graph family, and the simulation part of the paper is transparent and plausible. The main value would be in providing a tunable way to interpolate between purely local visibility and dense global connectivity. However, the ECoG-based conclusions are currently not supported by statistical inference, depend on unreported user-chosen thresholds, and rely on a crude power-law fitting procedure. These issues must be addressed before the paper can substantiate its headline claims about neural-state differences or about superior long-range dependency capture.

major comments (5)

[Sec. 2.3, Eq. (8)] The value of P0 used in the ECoG analysis is never reported. Since the binary adjacency matrix A_PVG is defined by thresholding P_ij at P0, and Fig. 2 shows that L, C, and k_max change sharply with P0, the rest/anesthesia differences in Fig. 3 may depend strongly on this choice. The authors should state the P0 value and provide a sensitivity analysis over P0 (e.g., a grid of P0 values in Fig. 3 or in a supplementary figure) to demonstrate that the reported ordering is stable.
[Sec. 2.3, Fig. 3] The claim that rest has "significantly higher" small-worldness than anesthesia is not supported by any statistical test. The shaded areas in Fig. 3 indicate standard deviations across 30 segments, but no paired test, permutation test, bootstrap confidence interval, or multiple-comparison correction is reported. Since the paper's central empirical claim is a difference between two brain states, inferential statistics are required to distinguish a stable effect from sampling variability.
[Sec. 2.3, Fig. 3(f)] The scale-free claim rests on a linear regression of the log-transformed degree distribution, but the fitting range (k_min), binning procedure, goodness-of-fit, confidence intervals, and comparison to alternative distributions (e.g., log-normal or stretched exponential) are omitted. The text's criterion gamma > 1 is too weak to identify scale-free behavior, and the more restrictive criterion 2 < gamma < 3 is not applied. The authors should report fitted exponents with uncertainty and use a more rigorous power-law testing procedure.
[Eq. (4)] The exponential decay P_ij = exp(-rho h_max) is introduced by analogy to quantum tunnelling and is not derived from the data or compared with alternative functional forms. Because the main empirical contrast is computed on graphs generated by this specific choice plus user-selected rho and P0, the authors should either provide a data-driven justification for the exponential form or demonstrate that the rest/anesthesia differences are robust to plausible alternative decay functions (e.g., power-law or linear decay). This is a concrete correctness-risk concern rather than a claim of circularity.
[Sec. 2.2] The claim that the PVG is superior to the classical VG in capturing long-range dependencies is based primarily on the maximum degree k_max, which trivially increases when the graph becomes denser. The comparison between k_max (up to 5000) and the autocorrelation-derived lag (3852) does not directly measure whether long-range dependencies in the time series are captured. The authors should validate the method on a ground-truth simulation with known long-range coupling, for example by computing the correlation between PVG connection probability and the true coupling strength.

minor comments (5)

[Fig. 3 caption and Sec. 2.3] The caption for Fig. 3 labels panel (b) as "Average degree," while the text in Sec. 2.3 refers to "network density." These are different quantities; the authors should make the terminology consistent.
[Sec. 2.3] In the sentence "Both rest and anesthesia conditions demonstrated small-world properties with rest showing significantly higher sigma than the rest condition," the phrase "than the rest condition" should read "than the anesthesia condition."
[Eqs. (7) and (8)] There is a potential inconsistency between the weighted matrix M_ij = W_ij * P_ij and the binary adjacency matrix A_PVG, which is defined by P_ij >= P0. If W_ij = 0 (e.g., equal values), then M_ij = 0 even though A_PVG,ij = 1. The authors should clarify which matrix is used for each reported network measure and whether edges with zero weight are included in the binary analysis.
[Fig. 3(f)] The axis/caption description says the power-law exponent gamma is shown "as a function of k," but the panels are plotted against rho. This should be corrected to "as a function of rho." Also, the formula in the text "P(k)~k−1" appears garbled and should be written as P(k) ~ k^{-gamma}.
[References and availability] The paper does not provide a data or code availability statement. Since the method is computationally simple, sharing code would substantially increase reproducibility; at minimum, the threshold choices and fitting details should be specified.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the PVG is defined by explicit equations, the network metrics are computed on external ECoG data, and none of the paper's load-bearing claims reduces to a fitted parameter or a self-citation chain.

full rationale

The derivation chain is self-contained. Equation (4) posits the exponential tunneling probability as a modeling ansatz, Eq. (8) defines the adjacency matrix from that probability plus a threshold, and the small-worldness and scale-free measures are then computed from the resulting graph on externally recorded ECoG data [19]. The power-law exponent is estimated by linear regression after graph construction, so the fit is descriptive rather than used to generate the graph. The authors' cited prior works ([14], [15], [20]) are contextual or concern preprocessing details; they are not invoked as uniqueness theorems or as the source of the central rest/anesthesia contrast. The main caveats—that P_ij's exponential form is an unvalidated ansatz and that the ECoG application does not report the P0 value or provide a sensitivity analysis—are robustness or reporting limitations, not circularity, because the reported differences are empirical consequences of the chosen construction rather than identities enforced by definition. No equation in the paper reduces the claimed prediction to a fitted parameter, a self-citation, or an equivalent restatement of the method's inputs.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim rests on the specific exponential decay rule (Eq. 4), the adoption of a prior weighted VG formula (Eq. 6), and the user-chosen parameters rho and P0. No new physical entities are introduced. The power-law exponent is a fitted descriptive quantity, not used to construct the graph.

free parameters (3)

rho (decay parameter)
Controls how quickly connection probability decays with obstruction height; results in Fig. 3 are reported across a range of rho, and the qualitative rest/anesthesia difference depends on this choice.
P0 (probability threshold) = varied (e.g., 0.5 in Fig. 1; other values in Fig. 2)
Binarizes the PVG; when P0=1 the graph reduces to the classical VG, so all downstream metrics depend on P0.
Power-law fitting range (k_min, binning)
The degree distribution exponent gamma is estimated by linear regression on log-transformed data; the choice of fitting range and binning is not specified and affects the scale-free conclusion.

assumptions (4)

ad hoc to paper The exponential form P_ij = exp(-rho h_max) models tunnelling probability.
Eq. (4) asserts this functional form without derivation from quantum mechanics or from data; it is the core modeling choice.
domain assumption The interaction strength W_ij = arctan((x_j - x_i)/(t_j - t_i)) from [13] is appropriate.
Eq. (6) adopts a prior weighted VG definition; the weighted matrix is not used in the binary analysis, so its role is unclear.
domain assumption Network measures (L, C, sigma, gamma) computed on the binarized PVG meaningfully characterize brain dynamics.
Section 2.3 assumes the standard graph metrics reflect physiological state without validating against ground truth or null models.
domain assumption The autocorrelation maximum lag of 3852 is a meaningful benchmark for long-range dependency.
Section 2.2 introduces this number without defining how it was computed.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Tunnelling Through Time Series: A Probabilistic Visibility Graph for Local and Global Pattern Discovery." pith.science (2026). https://pith.science/paper/U5RQMMLV

@misc{pith2026250701247,
  author       = {Pith},
  title        = {Pith review of: Tunnelling Through Time Series: A Probabilistic Visibility Graph for Local and Global Pattern Discovery},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/U5RQMMLV}},
  note         = {Machine review of arXiv:2507.01247}
}

read the original abstract

The growing availability of high-resolution, long-term time series data has highlighted the need for methods capable of capturing both local and global patterns. To address this, we introduce the Probabilistic Visibility Graph (PVG), a novel approach inspired by the quantum tunnelling phenomenon. The PVG extends the classical Visibility Graph (VG) by introducing probabilistic connections between time points that are obstructed in the VG due to intermediate values. We demonstrate the PVG's effectiveness in capturing long-range dependencies through simulations of amplitude-modulated signals and analysis of electrocorticography (ECoG) data under rest and anesthesia conditions. Key results show that the PVG presents distinct network properties between rest and anesthesia, with rest exhibiting stronger small-worldness and scale-free behavior, reflecting a hub-dominated, centralized connectivity structure, compared to anesthesia. These findings highlight the PVG's potential for analyzing complex signals with interacting temporal scales, offering new insights into neural dynamics and other real-world phenomena.

Figures

Figures reproduced from arXiv: 2507.01247 by the authors.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

22 extracted references · 12 canonical work pages

[1]

Long-Term Electroencephalographic Monitoring for Diagnosis and Management of Seizures,

T. D. Lagerlund, G. D. Cascino, K. M. Cicora, and F. W. Sharbrough, “Long-Term Electroencephalographic Monitoring for Diagnosis and Management of Seizures,” Mayo Clinic Proceedings, vol. 71, no. 10, pp. 1000–1006, Oct. 1996, doi: 10.4065/71.10.1000

work page doi:10.4065/71.10.1000 1996
[2]

FNSPID: A Comprehensive Financial News Dataset in Time Series,

Z. Dong, X. Fan, and Z. Peng, “FNSPID: A Comprehensive Financial News Dataset in Time Series,” in Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, in KDD ’24. New York, NY, USA: Association for Computing Machinery, Aug. 2024, pp. 4918–4927. doi: 10.1145/3637528.3671629

arXiv 2024
[3]

Long-term (1990–2019) monitoring of forest cover changes in the humid tropics,

C. Vancutsem et al., “Long-term (1990–2019) monitoring of forest cover changes in the humid tropics,” Science Advances, vol. 7, no. 10, p. eabe1603, Mar. 2021, doi: 10.1126/sciadv.abe1603

doi:10.1126/sciadv.abe1603 1990
[5]

DSTP-RNN: A dual-stage two-phase attention-based recurrent neural network for long-term and multivariate time series prediction,

Y. Liu, C. Gong, L. Yang, and Y. Chen, “DSTP-RNN: A dual-stage two-phase attention-based recurrent neural network for long-term and multivariate time series prediction,” Expert Systems with Applications, vol. 143, p. 113082, Apr. 2020, doi: 10.1016/j.eswa.2019.113082

work page arXiv 2020
[6]

Complex network approaches to nonlinear time series analysis,

Y. Zou, R. V. Donner, N. Marwan, J. F. Donges, and J. Kurths, “Complex network approaches to nonlinear time series analysis,” Physics Reports, vol. 787, pp. 1–97, Jan. 2019, doi: 10.1016/j.physrep.2018.10.005

doi:10.1016/j.physrep.2018.10.005 2019
[8]

Cross and joint ordinal partition transition networks for multivariate time series analysis,

H. Guo, J.-Y. Zhang, Y. Zou, and S.-G. Guan, “Cross and joint ordinal partition transition networks for multivariate time series analysis,” Front. Phys., vol. 13, no. 5, pp. 1–10, Oct. 2018, doi: 10.1007/s11467-018-0805-0

work page doi:10.1007/s11467-018-0805-0 2018
[9]

visible" to each other. Visibility is determined by the geometric criterion: 𝑥(𝑡

transforms a time series into a graph where each data point 𝑥(𝑡!) is represented as a node, and edges are drawn between nodes i and j if the corresponding data points are "visible" to each other. Visibility is determined by the geometric criterion: 𝑥(𝑡")<𝑥(𝑡!)+#$%!&'#(%")%!'%"(𝑡"−𝑡!),∀𝑡"∈,𝑡!,𝑡*- (1) This criterion ensures that a straight line connecting 𝑥...

work page 2022
[10]

From time series to complex networks: The visibility graph,

L. Lacasa, B. Luque, F. Ballesteros, J. Luque, and J. C. Nuño, “From time series to complex networks: The visibility graph,” Proceedings of the National Academy of Sciences, vol. 105, no. 13, pp. 4972–4975, Apr. 2008, doi: 10.1073/pnas.0709247105

doi:10.1073/pnas.0709247105 2008

Show all 22 references

[11]

Theta activity and cognitive functioning: Integrating evidence from resting-state and task-related developmental electroencephalography (EEG) research,

E. Tan et al., “Theta activity and cognitive functioning: Integrating evidence from resting-state and task-related developmental electroencephalography (EEG) research,” Dev Cogn Neurosci, vol. 67, p. 101404, Jun. 2024, doi: 10.1016/j.dcn.2024.101404

2024
[12]

The functional significance of delta oscillations in cognitive processing,

T. Harmony, “The functional significance of delta oscillations in cognitive processing,” Front. Integr. Neurosci., vol. 7, Dec. 2013, doi: 10.3389/fnint.2013.00083

2013
[13]

Weighted Visibility Graph With Complex Network Features in the Detection of Epilepsy,

S. Supriya, S. Siuly, H. Wang, J. Cao, and Y. Zhang, “Weighted Visibility Graph With Complex Network Features in the Detection of Epilepsy,” IEEE Access, vol. 4, pp. 6554–6566, 2016, doi: 10.1109/ACCESS.2016.2612242

2016
[14]

Seizure localization using pre ictal phase-amplitude coupling in intracranial electroencephalography,

N. E. Cámpora, C. J. Mininni, S. Kochen, and S. E. Lew, “Seizure localization using pre ictal phase-amplitude coupling in intracranial electroencephalography,” Sci Rep, vol. 9, no. 1, p. 20022, Dec. 2019, doi: 10.1038/s41598-019-56548-y

2019 doi
[16]

, WORLD SCIENTIFIC, 1995, pp. 441–445. doi: 10.1142/9789812833709_0030

1995 doi
[17]

Phase-Amplitude Coupling and Long-Range Phase Synchronization Reveal Frontotemporal Interactions during Visual Working Memory,

J. Daume, T. Gruber, A. K. Engel, and U. Friese, “Phase-Amplitude Coupling and Long-Range Phase Synchronization Reveal Frontotemporal Interactions during Visual Working Memory,” J. Neurosci., vol. 37, no. 2, pp. 313–322, Jan. 2017, doi: 10.1523/JNEUROSCI.2130-16.2016

2017 doi
[18]

Complex network measures of brain connectivity: Uses and interpretations,

M. Rubinov and O. Sporns, “Complex network measures of brain connectivity: Uses and interpretations,” NeuroImage, vol. 52, no. 3, pp. 1059–1069, Sep. 2010, doi: 10.1016/j.neuroimage.2009.10.003

2010 doi
[19]

A new method for quantifying the performance of EEG blind source separation algorithms by referencing a simultaneously recorded ECoG signal,

N. Oosugi, K. Kitajo, N. Hasegawa, Y. Nagasaka, K. Okanoya, and N. Fujii, “A new method for quantifying the performance of EEG blind source separation algorithms by referencing a simultaneously recorded ECoG signal,” Neural Networks, vol. 93, pp. 1–6, Sep. 2017, doi: 10.1016/j...

2017 doi
[20]

Estimating the Excitatory-Inhibitory Balance from Electrocorticography Data using Physics-Informed Neural Networks,

R. C. Sotero and J. Sanchez-Bornot, “Estimating the Excitatory-Inhibitory Balance from Electrocorticography Data using Physics-Informed Neural Networks,” in Proceedings of the 11th International Conference on Bioinformatics Research and Applications, in ICBRA ’24. New York, NY...

2025
[21]

Network ‘Small-World-Ness’: A Quantitative Method for Determining Canonical Network Equivalence,

M. D. Humphries and K. Gurney, “Network ‘Small-World-Ness’: A Quantitative Method for Determining Canonical Network Equivalence,” PLOS ONE, vol. 3, no. 4, p. e0002051, Apr. 2008, doi: 10.1371/journal.pone.0002051. 9

2008 doi
[22]

Statistical mechanics of complex networks,

R. Albert and A.-L. Barabási, “Statistical mechanics of complex networks,” Rev. Mod. Phys., vol. 74, no. 1, pp. 47–97, Jan. 2002, doi: 10.1103/RevModPhys.74.47

2002 doi
[23]

Scale-free networks are rare,

A. D. Broido and A. Clauset, “Scale-free networks are rare,” Nat Commun, vol. 10, no. 1, p. 1017, Mar. 2019, doi: 10.1038/s41467-019-08746-5

2019 doi
[24]

Evolution of networks,

S. N. Dorogovtsev and J. F. F. Mendes, “Evolution of networks,” Advances in Physics, vol. 51, no. 4, pp. 1079–1187, Jun. 2002, doi: 10.1080/00018730110112519

2002 doi
[25]

Power laws, Pareto distributions and Zipf’s law,

M. Newman, “Power laws, Pareto distributions and Zipf’s law,” Contemporary Physics, vol. 46, no. 5, pp. 323–351, Sep. 2005, doi: 10.1080/00107510500052444

2005 doi

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [1]

Long-Term Electroencephalographic Monitoring for Diagnosis and Management of Seizures,

T. D. Lagerlund, G. D. Cascino, K. M. Cicora, and F. W. Sharbrough, “Long-Term Electroencephalographic Monitoring for Diagnosis and Management of Seizures,” Mayo Clinic Proceedings, vol. 71, no. 10, pp. 1000–1006, Oct. 1996, doi: 10.4065/71.10.1000

work page doi:10.4065/71.10.1000 1996

[2] [2]

FNSPID: A Comprehensive Financial News Dataset in Time Series,

Z. Dong, X. Fan, and Z. Peng, “FNSPID: A Comprehensive Financial News Dataset in Time Series,” in Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, in KDD ’24. New York, NY, USA: Association for Computing Machinery, Aug. 2024, pp. 4918–4927. doi: 10.1145/3637528.3671629

arXiv 2024

[3] [3]

Long-term (1990–2019) monitoring of forest cover changes in the humid tropics,

C. Vancutsem et al., “Long-term (1990–2019) monitoring of forest cover changes in the humid tropics,” Science Advances, vol. 7, no. 10, p. eabe1603, Mar. 2021, doi: 10.1126/sciadv.abe1603

doi:10.1126/sciadv.abe1603 1990

[4] [5]

DSTP-RNN: A dual-stage two-phase attention-based recurrent neural network for long-term and multivariate time series prediction,

Y. Liu, C. Gong, L. Yang, and Y. Chen, “DSTP-RNN: A dual-stage two-phase attention-based recurrent neural network for long-term and multivariate time series prediction,” Expert Systems with Applications, vol. 143, p. 113082, Apr. 2020, doi: 10.1016/j.eswa.2019.113082

work page arXiv 2020

[5] [6]

Complex network approaches to nonlinear time series analysis,

Y. Zou, R. V. Donner, N. Marwan, J. F. Donges, and J. Kurths, “Complex network approaches to nonlinear time series analysis,” Physics Reports, vol. 787, pp. 1–97, Jan. 2019, doi: 10.1016/j.physrep.2018.10.005

doi:10.1016/j.physrep.2018.10.005 2019

[6] [8]

Cross and joint ordinal partition transition networks for multivariate time series analysis,

H. Guo, J.-Y. Zhang, Y. Zou, and S.-G. Guan, “Cross and joint ordinal partition transition networks for multivariate time series analysis,” Front. Phys., vol. 13, no. 5, pp. 1–10, Oct. 2018, doi: 10.1007/s11467-018-0805-0

work page doi:10.1007/s11467-018-0805-0 2018

[7] [9]

visible" to each other. Visibility is determined by the geometric criterion: 𝑥(𝑡

transforms a time series into a graph where each data point 𝑥(𝑡!) is represented as a node, and edges are drawn between nodes i and j if the corresponding data points are "visible" to each other. Visibility is determined by the geometric criterion: 𝑥(𝑡")<𝑥(𝑡!)+#$%!&'#(%")%!'%"(𝑡"−𝑡!),∀𝑡"∈,𝑡!,𝑡*- (1) This criterion ensures that a straight line connecting 𝑥...

work page 2022

[8] [10]

From time series to complex networks: The visibility graph,

L. Lacasa, B. Luque, F. Ballesteros, J. Luque, and J. C. Nuño, “From time series to complex networks: The visibility graph,” Proceedings of the National Academy of Sciences, vol. 105, no. 13, pp. 4972–4975, Apr. 2008, doi: 10.1073/pnas.0709247105

doi:10.1073/pnas.0709247105 2008