arxiv: 2605.03326 · v1 · submitted 2026-05-05 · 📊 stat.CO

Recognition: unknown

Sequential Bayesian Monitoring for Recoverable and Drifting Processes

Gordon J. Ross

Pith reviewed 2026-05-09 16:30 UTC · model grok-4.3

classification 📊 stat.CO

keywords Bayesian monitoringstatistical process controlrecoverable processessequential trackingposterior probabilityPhase II SPCdrifting parameters

0 comments

The pith

Bayesian recursions compute the probability a recoverable process is currently in control at any moment.

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

The paper develops monitoring methods for Phase II statistical process control that focus on whether a process is acceptable right now, rather than whether it has ever changed in the past. For processes that can switch between in-control and out-of-control states and recover, it derives recursive formulas to update the posterior probability of current in-control status. For cases where a hidden parameter drifts over time, it instead tracks the posterior probability that the parameter stays inside a preset acceptable region. This approach supports continued monitoring after signals occur or after corrective actions are taken, situations common in ongoing quality control.

Core claim

For recoverable processes that alternate between in-control and out-of-control states according to a probabilistic switching model, recursive updates give the exact posterior probability that the process is in control at the current time step. For sequential tracking where a latent parameter evolves, similar recursions monitor the posterior probability that the parameter lies inside an acceptable region. These procedures are demonstrated on time-between-failure data, Gaussian and binomial tracking examples, and a multivariate held-out dataset of white wine quality measurements.

What carries the argument

Recursive Bayesian updates for the posterior probability of current in-control status (under Markovian state alternation) or of parameter acceptability (under time-evolving latent models).

If this is right

Monitoring can continue without reset after an out-of-control signal if corrective action restores the process.
The same framework applies to both discrete state-switching and continuous drifting-parameter problems.
The methods produce calibrated probabilities that can be used directly as decision thresholds in ongoing surveillance.

Where Pith is reading between the lines

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

The recursions could be embedded in real-time dashboards that display live probability of acceptability rather than binary alarms.
Extensions to non-Markovian recovery patterns would require new derivation steps but might retain the same monitoring goal.
Comparison against frequentist control charts on the same held-out data would clarify when the Bayesian current-state focus changes decisions.

Load-bearing premise

The processes must follow the assumed probabilistic models for state switching or parameter evolution so that the posterior probabilities can be updated recursively without approximation.

What would settle it

Run the derived recursions on a simple simulated recoverable process with known state sequence and compare the reported posterior probabilities against exact posteriors obtained by direct enumeration; any systematic mismatch would show the recursions are incorrect.

Figures

Figures reproduced from arXiv: 2605.03326 by Gordon J. Ross.

**Figure 1.** Figure 1: Sequential monitoring of a Bernoulli proportion that switches between the in-control state view at source ↗

**Figure 2.** Figure 2: Static diagnostic for the wine quality illustration. The plot shows squared Mahalanobis view at source ↗

**Figure 3.** Figure 3: Representative posterior acceptability path for the wine pseudo-monitoring illustration. view at source ↗

read the original abstract

In many Phase II statistical process control (SPC) problems, the main concern is not whether a monitored process has ever changed, but whether it is currently operating at an acceptable level. This distinction is especially important when monitoring continues after a signal, or when corrective action may restore the process. We develop Bayesian monitoring procedures for this formulation of the Phase II task. For recoverable processes that may alternate between in-control and out-of-control states, we derive recursions for the posterior probability that the process is presently in control. For sequential tracking problems in which a latent parameter evolves over time, we monitor the posterior probability that the parameter lies inside an acceptable region of behavior. The methods are studied through calibrated time-between-failure experiments, Gaussian and Binomial tracking examples, and a held-out multivariate data illustration using white wine quality measurements.

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.

Desk Editor's Note private letter to a colleague

Ross derives recursive Bayesian updates for the current in-control probability in recoverable processes and acceptable-region probability for drifting parameters, a practical reframing of Phase II SPC that works under its stated models but assumes known transition rates.

read the letter

The main takeaway is that this paper gives explicit recursions for the posterior probability the process is in control right now, for cases where it can switch back and forth, plus a similar posterior for a drifting parameter staying inside an acceptable band. That focus on present state rather than any past change is a useful shift for monitoring setups that allow recovery or correction. The derivations sit on standard Bayesian updating under Markov switching and latent evolution, and the abstract shows they are tested on time-between-failure calibrations, Gaussian and binomial tracking, and a held-out multivariate wine quality example. Those elements are the actual new pieces and they are presented cleanly without extra machinery. The soft spot is the dependence on fixed, known transition probabilities for the state switches and drifts. If those rates are misspecified or the process has memory beyond Markov, the recursions no longer give the correct posterior, and nothing in the abstract indicates sensitivity checks or data-driven estimation of the transitions. That assumption is central, so it limits how far the methods travel without further work. This is for readers already working in Bayesian statistical process control or sequential monitoring who need tools for ongoing rather than retrospective detection. The recursions and examples give enough concrete value that a serious referee should see it, with the review focused on robustness to the transition parameters and any gaps in the full derivations.

Referee Report

2 major / 2 minor

Summary. The manuscript develops Bayesian monitoring procedures for Phase II statistical process control, emphasizing whether a process is currently operating acceptably rather than whether it has ever changed. For recoverable processes that alternate between in-control and out-of-control states, it derives recursions for the posterior probability that the process is presently in control. For sequential tracking of drifting latent parameters, it monitors the posterior probability that the parameter lies inside an acceptable region. The methods are illustrated through calibrated time-between-failure experiments, Gaussian and Binomial tracking examples, and a held-out multivariate illustration using white wine quality data.

Significance. If the recursive derivations hold under the assumed models, the work provides a targeted Bayesian tool for ongoing Phase II monitoring that accommodates recovery and drift, addressing a practical distinction from standard change detection. The calibrated experiments and real-data example add credibility to the sequential updating approach, though overall significance hinges on the validity of the Markovian and evolution assumptions for the target applications.

major comments (2)

[§3] §3 (Recoverable Processes): The recursions for the in-control posterior probability assume a Markov chain with fixed, known transition probabilities between states. The manuscript provides no procedure for estimating these probabilities from data nor any robustness checks against misspecification (e.g., non-Markovian alternation or unknown rates), which directly affects whether the posterior is correctly computed for general recoverable processes.
[§4] §4 (Drifting Parameters): The monitoring rule for the posterior probability that the latent parameter lies in the acceptable region depends on a specific parametric evolution model for the drift. Without explicit validation that the acceptable region definition and evolution model match the Gaussian/Binomial examples, the claim that this yields reliable sequential tracking is not fully supported by the reported experiments.

minor comments (2)

[Abstract] The abstract refers to 'calibrated time-between-failure experiments' without stating the calibration metric or acceptance criteria used to tune the procedures.
[Introduction] Notation for the state probabilities and acceptable regions could be introduced with a single summary table to improve readability across the recoverable and drifting cases.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the thoughtful and constructive comments on our manuscript. We address each major comment in detail below, providing clarifications on the scope of our contributions while acknowledging areas where additional discussion can strengthen the presentation.

read point-by-point responses

Referee: [§3] §3 (Recoverable Processes): The recursions for the in-control posterior probability assume a Markov chain with fixed, known transition probabilities between states. The manuscript provides no procedure for estimating these probabilities from data nor any robustness checks against misspecification (e.g., non-Markovian alternation or unknown rates), which directly affects whether the posterior is correctly computed for general recoverable processes.

Authors: The recursions are indeed derived under a Markovian model with fixed, known transition probabilities, which are treated as hyperparameters of the monitoring procedure. The primary contribution of this section is the exact recursive computation of the posterior probability that the process is currently in control, given the model. We do not provide an estimation procedure because the focus is on the sequential updating rule for Phase II monitoring rather than on model fitting; in applications, the transition probabilities can be specified from domain expertise or estimated in a separate step from historical data using standard HMM techniques such as the Baum-Welch algorithm or Bayesian posterior sampling. We agree that robustness to misspecification (including potential non-Markovian behavior) is a relevant practical concern. In revision we will add a brief discussion paragraph noting this modeling assumption and outlining possible estimation approaches, while clarifying that the derived recursions are exact under the stated model. revision: partial
Referee: [§4] §4 (Drifting Parameters): The monitoring rule for the posterior probability that the latent parameter lies in the acceptable region depends on a specific parametric evolution model for the drift. Without explicit validation that the acceptable region definition and evolution model match the Gaussian/Binomial examples, the claim that this yields reliable sequential tracking is not fully supported by the reported experiments.

Authors: The Gaussian and Binomial examples are constructed directly from the parametric evolution models used in the derivation (random-walk or linear-Gaussian dynamics for the latent parameter in the Gaussian case, and the corresponding binomial state evolution). The acceptable regions are defined to be consistent with the process characteristics and parameter ranges in each example, as described in the respective subsections. The experiments therefore illustrate the exact sequential monitoring rule under correctly specified models, which is the intended demonstration. We acknowledge that the manuscript could state this alignment more explicitly. In revision we will add clarifying sentences in §4 and the example sections to confirm that the evolution models and acceptable-region definitions match the simulated data-generating processes, thereby strengthening the support for the reported tracking performance. revision: yes

Circularity Check

0 steps flagged

Standard Bayesian recursions for HMM-style monitoring; no reduction to inputs by construction

full rationale

The paper derives recursive updates for the posterior probability that a recoverable process is currently in-control (or that a drifting parameter lies in an acceptable region) by applying Bayes' rule to an assumed Markovian state process with known transition probabilities. This is a direct, self-contained application of the forward algorithm for hidden Markov models and does not involve self-definition, fitted inputs renamed as predictions, or load-bearing self-citations. The central claims rest on the explicit model assumptions stated in the abstract and methods, which are externally verifiable and not equivalent to the derived recursions by construction. No circular steps are present.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Only the abstract is available; full details on parameters, assumptions, and derivations are inaccessible. The approach implicitly relies on standard Bayesian updating and unspecified models for state transitions and parameter dynamics.

axioms (1)

standard math Bayesian updating can be applied recursively to compute current-state posteriors for the described process models.
The core contribution is the derivation of such recursions.

pith-pipeline@v0.9.0 · 5426 in / 1037 out tokens · 51140 ms · 2026-05-09T16:30:42.465704+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

300 extracted references · 75 canonical work pages

[1]

Journal of Quality Technology , year =

Bourazas, Konstantinos and Kiagias, Dimitrios and Tsiamyrtzis, Panagiotis , title =. Journal of Quality Technology , year =
[2]

Predictive ratio

Bourazas, Konstantinos and Sobas, Fr. Predictive ratio. Journal of Quality Technology , year =
[3]

Design and properties of the predictive ratio

Bourazas, Konstantinos and Sobas, Fr. Design and properties of the predictive ratio. Journal of Quality Technology , year =
[4]

Quality and Reliability Engineering International , year =

Ali, Sajid , title =. Quality and Reliability Engineering International , year =
[5]

Journal of Quality Technology , year =

Capizzi, Giovanna and Masarotto, Guido , title =. Journal of Quality Technology , year =
[6]

Quality and Reliability Engineering International , year =

Wang, Cheng and Castagliola, Philippe and Guo, Baocai , title =. Quality and Reliability Engineering International , year =. doi:10.1002/qre.70222 , note =

work page doi:10.1002/qre.70222
[7]

Adams, Ryan Prescott and MacKay, David J. C. , title =. arXiv preprint arXiv:0710.3742 , year =

work page Pith review arXiv
[8]

Language Resources and Evaluation , author =

Authorship attribution in the wild , volume =. Language Resources and Evaluation , author =. 2011 , pages =. doi:10.1007/s10579-009-9111-2 , abstract =

work page doi:10.1007/s10579-009-9111-2 2011
[9]

The ASA’ s Statement on p-Values: Context, Process, and Purpose.The American Statisti- cian2016;70(2):129–133

A. The American Statistician , author =. 2016 , pages =. doi:10.1080/00031305.2016.1148630 , language =

work page doi:10.1080/00031305.2016.1148630 2016
[10]

Journal of the Association for Information Science and Technology , author =

Determining if two documents are written by the same author , volume =. Journal of the Association for Information Science and Technology , author =. 2014 , keywords =. doi:10.1002/asi.22954 , abstract =

work page doi:10.1002/asi.22954 2014
[11]

Narayanan, Arvind and Paskov, Hristo and Gong, Neil Zhenqiang and Bethencourt, John and Stefanov, Emil and Shin, Eui Chul Richard and Song, Dawn , month = may, year =. On the. 2012. doi:10.1109/SP.2012.46 , abstract =

work page doi:10.1109/sp.2012.46 2012
[12]

Xun, Guangxu and Li, Yaliang and Zhao, Wayne Xin and Gao, Jing and Zhang, Aidong , pages =. A
[13]

Advances in Neural Information Processing Systems 26 , author =

Scalable. Advances in Neural Information Processing Systems 26 , author =. 2013 , pages =

2013
[14]

Advances in Neural Information Processing Systems 28 , author =

Dependent. Advances in Neural Information Processing Systems 28 , author =. 2015 , pages =

2015
[15]

Literary and Linguistic Computing , author =

Quantitative. Literary and Linguistic Computing , author =. 2007 , pages =. doi:10.1093/llc/fqm020 , abstract =

work page doi:10.1093/llc/fqm020 2007
[16]

Function

Kestemont, Mike , month = apr, year =. Function. Proceedings of the 3rd
[17]

and Lewis, Er Genkin David D

Madigan, David and Genkin, Alexander and Lewis, David D. and Lewis, Er Genkin David D. and Argamon, Shlomo and Fradkin, Dmitriy and Ye, Li and Consulting, David D. Lewis , year =. Author. In
[18]

Literary and Linguistic Computing , author =

Ngram and. Literary and Linguistic Computing , author =. 2003 , pages =. doi:10.1093/llc/18.4.423 , abstract =

work page doi:10.1093/llc/18.4.423 2003
[19]

Physical Review Letters , author =

Universality in. Physical Review Letters , author =. 2006 , pages =. doi:10.1103/PhysRevLett.96.051102 , abstract =

work page doi:10.1103/physrevlett.96.051102 2006
[20]

Journal of Geophysical Research: Space Physics , author =

The. Journal of Geophysical Research: Space Physics , author =. 2018 , keywords =. doi:10.1029/2018JA025740 , abstract =

work page doi:10.1029/2018ja025740 2018
[21]

Physical Review E , author =

Stationarity of extreme bursts in the solar wind , volume =. Physical Review E , author =. 2014 , pages =. doi:10.1103/PhysRevE.89.052812 , abstract =

work page doi:10.1103/physreve.89.052812 2014
[22]

Physical Review E , author =

Waiting-time distributions of magnetic discontinuities:. Physical Review E , author =. 2009 , pages =. doi:10.1103/PhysRevE.80.046401 , abstract =

work page doi:10.1103/physreve.80.046401 2009
[23]

The Astrophysical Journal Letters , author =

Solar. The Astrophysical Journal Letters , author =. 2001 , pages =. doi:10.1086/323178 , abstract =

work page doi:10.1086/323178 2001
[24]

The Astrophysical Journal , author =

Reconciliation of. The Astrophysical Journal , author =. 2010 , pages =. doi:10.1088/0004-637X/717/2/683 , abstract =

work page doi:10.1088/0004-637x/717/2/683 2010
[25]

, keywords =

A. The Astrophysical Journal , author =. 2015 , pages =. doi:10.1088/0004-637X/801/1/57 , language =

work page doi:10.1088/0004-637x/801/1/57 2015
[26]

The Astrophysical Journal , author =

The. The Astrophysical Journal , author =. 2000 , pages =. doi:10.1086/312739 , abstract =

work page doi:10.1086/312739 2000
[27]

Chinese Journal of Astronomy and Astrophysics , author =

Waiting. Chinese Journal of Astronomy and Astrophysics , author =. 2005 , pages =. doi:10.1088/1009-9271/5/2/010 , abstract =

work page doi:10.1088/1009-9271/5/2/010 2005
[28]

Astronomy & Astrophysics , author =

Different triggering mechanisms for solar flares and coronal mass ejections , volume =. Astronomy & Astrophysics , author =. 2008 , pages =. doi:10.1051/0004-6361:200810164 , abstract =

work page doi:10.1051/0004-6361:200810164 2008
[29]

The Astrophysical Journal Letters , author =

Waiting. The Astrophysical Journal Letters , author =. 2014 , pages =. doi:10.1088/2041-8205/792/2/L26 , abstract =

work page doi:10.1088/2041-8205/792/2/l26 2014
[30]

Monthly Notices of the Royal Astronomical Society: Letters , author =

Waiting time distributions of solar and stellar flares:. Monthly Notices of the Royal Astronomical Society: Letters , author =. 2018 , pages =. doi:10.1093/mnrasl/sly117 , abstract =

work page doi:10.1093/mnrasl/sly117 2018
[31]

AIP Conference Proceedings , author =

Statistical properties of solar flares and coronal mass ejections through the solar cycle , volume =. AIP Conference Proceedings , author =. 2016 , pages =. doi:10.1063/1.4943806 , number =

work page doi:10.1063/1.4943806 2016
[32]

The Astrophysical Journal Letters , author =

Stochasticity and. The Astrophysical Journal Letters , author =. 2014 , pages =. doi:10.1088/2041-8205/781/1/L1 , abstract =

work page doi:10.1088/2041-8205/781/1/l1 2014
[33]

Astrophysics and Space Science , author =

Waiting. Astrophysics and Space Science , author =. 2006 , keywords =. doi:10.1007/s10509-005-9036-1 , abstract =

work page doi:10.1007/s10509-005-9036-1 2006
[34]

Similarities between extreme events in the solar-terrestrial system by means of nonextensivity , volume =. Nonlin. Processes Geophys. , author =. 2011 , pages =. doi:10.5194/npg-18-563-2011 , abstract =

work page doi:10.5194/npg-18-563-2011 2011
[35]

Physical Review E , author =

Return interval distribution of extreme events and long-term memory , volume =. Physical Review E , author =. 2008 , pages =. doi:10.1103/PhysRevE.78.051113 , abstract =

work page doi:10.1103/physreve.78.051113 2008
[36]

Solar Physics , author =

The. Solar Physics , author =. 2003 , keywords =. doi:10.1023/A:1024222511574 , abstract =

work page doi:10.1023/a:1024222511574 2003
[37]

Shalizi, and M

Power-. SIAM Review , author =. 2009 , pages =. doi:10.1137/070710111 , abstract =

work page doi:10.1137/070710111 2009
[38]

Journal of Applied Probability , author =

A. Journal of Applied Probability , author =. 1974 , pages =

1974
[39]

Journal of Geophysical Research: Solid Earth , author =

Triggering of seismicity at short timescales following. Journal of Geophysical Research: Solid Earth , author =. 2003 , keywords =. doi:10.1029/2002JB001946 , abstract =

work page doi:10.1029/2002jb001946 2003
[40]

Physical Review E , author =

Hawkes process model with a time-dependent background rate and its application to high-frequency financial data , volume =. Physical Review E , author =. 2017 , pages =. doi:10.1103/PhysRevE.96.012303 , abstract =

work page doi:10.1103/physreve.96.012303 2017
[41]

Physical Review E , author =

Quantifying reflexivity in financial markets:. Physical Review E , author =. 2012 , pages =. doi:10.1103/PhysRevE.85.056108 , abstract =

work page doi:10.1103/physreve.85.056108 2012
[42]

Valkeila, Esko , month = feb, year =. An. doi:10.1111/j.1751-5823.2008.00054_18.x , journal =

work page doi:10.1111/j.1751-5823.2008.00054_18.x 2008
[43]

Journal of the American Statistical Association , author =

Optimal. Journal of the American Statistical Association , author =. 2012 , keywords =

2012
[44]

D., Wilkinson, M

Information criteria for astrophysical model selection , volume =. Monthly Notices of the Royal Astronomical Society: Letters , author =. 2007 , keywords =. doi:10.1111/j.1745-3933.2007.00306.x , abstract =

work page doi:10.1111/j.1745-3933.2007.00306.x 2007
[45]

The Annals of Statistics7(1), 1–26 (1979) https://doi.org/10.1214/aos/1176344552

Estimating the. The Annals of Statistics , author =. 1978 , mrnumber =. doi:10.1214/aos/1176344136 , abstract =

work page doi:10.1214/aos/1176344136 1978
[46]

Systematic Biology , author =

Model. Systematic Biology , author =. 2004 , pages =. doi:10.1080/10635150490522304 , abstract =

work page doi:10.1080/10635150490522304 2004
[47]

Space Weather , author =

On the probability of occurrence of extreme space weather events , volume =. Space Weather , author =. 2012 , keywords =. doi:10.1029/2011SW000734 , abstract =

work page doi:10.1029/2011sw000734 2012
[48]

Space Weather , author =

Using. Space Weather , author =. 2018 , keywords =. doi:10.1002/2017SW001727 , abstract =

work page doi:10.1002/2017sw001727 2018
[49]

Bootstrap confidence intervals,

Bootstrap confidence intervals , volume =. Statistical Science , author =. 1996 , mrnumber =. doi:10.1214/ss/1032280214 , abstract =

work page doi:10.1214/ss/1032280214 1996
[50]

Risk Analysis , author =

The. Risk Analysis , author =. 2017 , keywords =. doi:10.1111/risa.12765 , abstract =

work page doi:10.1111/risa.12765 2017
[51]

N., Daly , E., Daglis , I., Kappenman , J

Effects of. Space Weather , author =. 2004 , file =. doi:10.1029/2003SW000044 , abstract =

work page doi:10.1029/2003sw000044 2004
[52]

Bulletin of the Seismological Society of America , author =

Explanation of temporal clustering of tsunami sources using the epidemic-type aftershock sequence model , volume =. Bulletin of the Seismological Society of America , author =. 2014 , note =. doi:10.1785/0120130275 , abstract =

work page doi:10.1785/0120130275 2014
[53]

The Annals of Applied Statistics , author =

Modeling for seasonal marked point processes:. The Annals of Applied Statistics , author =. 2015 , mrnumber =. doi:10.1214/14-AOAS796 , abstract =

work page doi:10.1214/14-aoas796 2015
[54]

Science , author =

Length of the. Science , author =. 1991 , pmid =. doi:10.1126/science.254.5032.698 , abstract =

work page doi:10.1126/science.254.5032.698 1991
[55]

Physical Review Letters , author =

Intensity. Physical Review Letters , author =. 2006 , pages =. doi:10.1103/PhysRevLett.96.051103 , abstract =

work page doi:10.1103/physrevlett.96.051103 2006
[56]

Space Weather , author =

A statistical solar flare forecast method:. Space Weather , author =. 2005 , pages =. doi:10.1029/2004SW000131 , language =

work page doi:10.1029/2004sw000131 2005
[57]

S., Higgins , P

Toward. The Astrophysical Journal Letters , author =. 2012 , pages =. doi:10.1088/2041-8205/747/2/L41 , abstract =

work page doi:10.1088/2041-8205/747/2/l41 2012
[58]

Solar Physics , author =

Solar. Solar Physics , author =. 2013 , keywords =. doi:10.1007/s11207-011-9896-1 , abstract =

work page doi:10.1007/s11207-011-9896-1 2013
[59]

Solar Physics , author =

Estimating the. Solar Physics , author =
[60]

Solar Physics , author =

Predicting. Solar Physics , author =
[61]

Journal of Physics of the Earth , author =

The centenary of the omori formula for a decay law of aftershock activity , volume =. Journal of Physics of the Earth , author =. 1995 , pages =. doi:10.4294/jpe1952.43.1 , abstract =

work page doi:10.4294/jpe1952.43.1 1995
[62]

Journal of Geophysical Research: Solid Earth , author =

Bayesian analysis of the modified. Journal of Geophysical Research: Solid Earth , author =. 2012 , keywords =. doi:10.1029/2011JB009054 , abstract =

work page doi:10.1029/2011jb009054 2012
[63]

1956 , pages =

Annals of Geophysics , author =. 1956 , pages =. doi:10.4401/ag-5590 , abstract =

work page doi:10.4401/ag-5590 1956
[64]

Bulletin of the Seismological Society of America , author =

Applying a. Bulletin of the Seismological Society of America , author =. 2007 , pages =. doi:10.1785/0120060181 , language =

work page doi:10.1785/0120060181 2007
[65]

Annals of Applied Statistics , author =

Spatially inhomogeneous background rate estimators and uncertainty quantification for nonparametric. Annals of Applied Statistics , author =. 2016 , mrnumber =

2016
[66]

Journal of the American Statistical Association , author =

Estimation of. Journal of the American Statistical Association , author =. 2008 , keywords =. doi:10.1198/016214508000000148 , abstract =

work page doi:10.1198/016214508000000148 2008
[67]

Annals of the Institute of Statistical Mathematics , author =

Space-. Annals of the Institute of Statistical Mathematics , author =. 1998 , pages =. doi:10.1023/A:1003403601725 , abstract =

work page doi:10.1023/a:1003403601725 1998
[68]

Science (New York, N.Y.) , author =

Earthquake hazard after a mainshock in california , volume =. Science (New York, N.Y.) , author =. 1989 , pmid =. doi:10.1126/science.243.4895.1173 , abstract =

work page doi:10.1126/science.243.4895.1173 1989
[69]

Nature , author =

Real-time forecasts of tomorrow's earthquakes in. Nature , author =. 2005 , pmid =. doi:10.1038/nature03622 , abstract =

work page doi:10.1038/nature03622 2005
[70]

Scientific Reports , author =

Forecasting large aftershocks within one day after the main shock , volume =. Scientific Reports , author =. 2013 , pages =. doi:10.1038/srep02218 , language =

work page doi:10.1038/srep02218 2013
[71]

Geophysical Research Letters , author =

Immediate and updated forecasting of aftershock hazard , volume =. Geophysical Research Letters , author =. 2006 , file =. doi:10.1029/2006GL025888 , abstract =

work page doi:10.1029/2006gl025888 2006
[72]

Geosciences , author =

Post. Geosciences , author =. 2019 , keywords =. doi:10.3390/geosciences9080355 , abstract =

work page doi:10.3390/geosciences9080355 2019
[73]

Bulletin of the Seismological Society of America , author =

On the estimation of seismic detection thresholds , volume =. Bulletin of the Seismological Society of America , author =. 1975 , pages =

1975
[74]

Spatial Statistics , author =

Handling missing data in self-exciting point process models , volume =. Spatial Statistics , author =. 2019 , pages =. doi:10.1016/j.spasta.2018.12.004 , abstract =

work page doi:10.1016/j.spasta.2018.12.004 2019
[75]

Journal of Geophysical Research: Solid Earth , author =

Importance of small earthquakes for stress transfers and earthquake triggering , volume =. Journal of Geophysical Research: Solid Earth , author =. 2005 , keywords =. doi:10.1029/2004JB003286 , abstract =

work page doi:10.1029/2004jb003286 2005
[76]

Technical Report , author =

Bayesian. Technical Report , author =
[77]

Assessing the

Alaiz-Rodriguez, Rocio and Japkowicz, Nathalie , editor =. Assessing the. Canadian. 2008 , keywords =

2008
[78]

Journal of the Royal Statistical Society Series A , author =

The synoptic problem: on. Journal of the Royal Statistical Society Series A , author =. 2012 , pages =

2012
[79]

ACM Transactions on Information Systems , author =

Writeprints:. ACM Transactions on Information Systems , author =
[80]

Journal of Data Science , author =

A. Journal of Data Science , author =. 2007 , keywords =

2007

Showing first 80 references.