Exact Time Research Articles

Revisiting the trading activity of high-frequency trading firms around ultra-fast flash events

We investigate high-frequency traders’ behavior in the context of the fastest and most extreme price movements (EPMs) that can be observed in the market, specifically ultra-fast flash events, challenging the methodologies employed in the academic and practitioner literature for identifying sudden liquidity black holes. To refine the price-shock identification methodology, we introduce a new approach called sequence-based flash events (SFEs), which relies on tick sequences instead of predetermined fixed-time intervals within which all flash events in the sample are assumed to occur. This alternative methodology offers the advantage of pinpointing the exact time and duration of a crash, which, in turn, provides a way to more accurately define the observation windows around it. We compare our sample of SFEs with both the so-called “mini flash crashes”, as identified by the Nanex detection algorithm, and the so-called EPMs, as identified by Brogaard et al. (2018). We use close and open prices, as well as high and low prices. Based on our sample of SFEs, we find no evidence that HFTs trigger extreme price shocks. However, we find that HFTs exacerbate SFEs by increasing the net imbalance in the direction of these shocks as they occur. Finally, we show that the choice of the price-shock identification methodology is critical. Thus, we urge regulators to exercise caution and avoid hasty conclusions regarding HFTs’ contribution to price stability in stressful market conditions.

The earliest development of the Vilnius Franciscan Church and monastery area

When the German Town (a settlement of Catholic merchants and craftsmen) began to emerge in the western part of the present Old Town of Vilnius in the last decades of the 14th century, the Catholic community relocated to this area. St Nicholas Church was built here before 1387 and it was the first Catholic church in the Old Town, around which the German City was established. However, in addition to this house of worship, there was another in the Catholic part of the city – the Franciscan Church of the Assumption of the Blessed Virgin Maria and a monastery.This monastery is considered to be the oldest one in Vilnius, however, the exact time of its construction, as well as that of the church, has not been determined. Due to the lack of historical sources and precise dates, the construction has been dated by different researchers from the first half of the 14th century to the beginning of the 15th century, i.e., with an error of 100 years. Archaeological investigations in the vicinity of the church and the monastery, which have been carried out since 1993, have provided some new hypotheses concerning the chronology and the earliest nature of human activity in this territory. The most recent excavations took place here in 2023, providing new insights into the development of the area through the discovery of objects and interdisciplinary research (radiocarbon dating, analysis of archaeobotanical remains, architectural research). This paper discusses the chronology and nature of the earliest development of the Franciscan church and monastery area, based on the available historical sources, archaeological and architectural research, including the results of most recent investigations.

Investigating eating architecture and the impact of the precision of recorded eating time: a cross-sectional study.

The precision of recorded eating times directly affects the estimation of eating architecture i.e. size, timing, and frequency of eating. The impact of imprecise timing on estimates and associations of eating architecture with health remains unclear. We compared eating architecture variables derived from precise versus broad timing methods and examined associations with anthropometric and diet-related outcomes. Cross-sectional data came from 3-day diet diaries of 7-year-old children in the Avon Longitudinal Study of Parents and Children. We derived average size, timing, and frequency of eating, utilizing exact times (precise, n=4855) and mid-point meal slot times (broad, n=7285). Intraclass correlation coefficients (ICC) estimated agreement between methods. Bland-Altman analysis determined mean difference and limits of agreement (LOA). Correlations (95% confidence intervals) estimated associations between eating architecture variables and anthropometric or diet-related traits. Agreement varied from moderate to excellent for size (ICC 0.75), last or first time (ICC 0.80 or 0.58), and frequency (ICC 0.43) of Eating Occasions (EOs). Broad times underestimated eating frequency (2.2 times/day; LOA -1, 5) and overestimated size (83g; LOA -179, 13), last time (50min; LOA -142, 42), inter-meal intervals (68min; LOA -126, -11) and eating window (49min; LOA -161, 63). Directions of eating architecture intercorrelations were consistent regardless of time precision but varied in magnitude, e.g., larger EO size correlated with lower eating frequency but was stronger with precise time (rprecise=-0.54 (95% CI -0.56, -0.52); rbroad=-0.24 (-0.27, -0.22)). Correlations with anthropometric and diet outcomes were also directionally consistent. Precise timing improves the estimation of eating architecture. Differences in estimation will affect descriptions of children's eating habits and possibly dietary guidance. However, consistent directional associations across timing methods, suggest that broad times could provide a pragmatic method for investigating eating architecture associations in large samples.

Time-variant quadratic programming solving by using finitely-activated RNN models with exact settling time

Time-variant quadratic programming solving by using finitely-activated RNN models with exact settling time

Classification-augmented survival estimation (CASE): A novel method for individualized long-term survival prediction with application to liver transplantation.

Survival analysis is critical in many fields, particularly in healthcare where it can guide medical decisions. Conventional survival analysis methods like Kaplan-Meier and Cox proportional hazards models to generate survival curves indicating probability of survival v. time have limitations, especially for long-term prediction, due to assumptions that all instances follow a general population-level survival curve. Machine learning classification models, even those designed for survival predictions like random survival forest (RSF), also struggle to provide accurate long-term predictions due to class imbalance. We improve upon traditional survival machine learning approaches through a novel framework called classification-augmented survival estimation (CASE), which treats survival as a classification task that ultimately yields survival curves, beginning with dataset augmentation to improve class imbalance for use with any classification model. Unlike other approaches, CASE additionally provides an exact survival time prediction. We demonstrate CASE on a liver transplant case study to predict >20 years survival post-transplant, finding that CASE dataset augmentation improved AUCs from 0.69 to 0.88 and F1 scores from 0.32 to 0.73. Compared to Kaplan-Meier, Cox, and RSF survival models, the CASE framework demonstrated better performance across various existing survival metrics, as well as our novel metric, mean of individual areas under the survival curve (mAUSC). Further, we develop novel temporal feature importance methods to understand how different features may vary in survival importance over time, potentially providing actionable insights in real-world survival problems.

In search of lost time: Discrete- versus continuous-time models of working alliance and symptom severity.

The therapeutic alliance is one of the most stable predictors of symptom burden over the course of therapy. So far, this effect has only been examined on the basis of sessions. Continuous-time models (CTM) allow this relationship to be modeled as a continuous process in which the actual time interval between measurements is considered. The aim of the present study was to compare the fit of discrete-time models (DTM) of the alliance-symptom relationship with CTM using different time variables (sessions vs. actual time interval). Data from 1,413 patients at a university psychotherapy outpatient clinic were analyzed. The alliance and symptom burden were assessed each session with the Bernese Session Report and the Hopkins Symptom Checklist-Short-Form, respectively. Different DTM and CTM were estimated using the R-package ctsem and compared in their fit via the Akaike information criterion. CTMs with session as the time unit fitted the data best. Significant negative within-person effects of alliance and symptom burden were found. These effects showed a significant positive correlation, implying that individuals with a stronger effect of the alliance on symptom severity also showed a stronger effect of symptom severity on the alliance. When modeling the relationship of symptom severity and alliance, it seems to be of more importance to capture the fact that a session occurred than to capture the exact time intervals between sessions. Future studies should examine this finding for other psychotherapeutic factors. Interpersonal factors might explain the positive association of the reciprocal alliance-symptom effects. (PsycInfo Database Record (c) 2025 APA, all rights reserved).

Occurrence and Time of Onset of Intraventricular Hemorrhage in Preterm Neonates

Intraventricular hemorrhage (IVH) has been described to typically occur during the early hours of life (HOL); however, the exact time of onset is still unknown. To investigate the temporal distribution of IVH reported in very preterm neonates. PubMed, Embase, Cochrane Library, and Web of Science were searched on May 9, 2024. Articles were selected in which at least 2 cranial ultrasonographic examinations were performed in the first week of life to diagnose IVH. Studies with only outborn preterm neonates were excluded. Data were extracted independently by 3 reviewers. A random-effects model was applied. This study is reported following the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guideline. The Quality in Prognostic Studies 2 tool was used to assess the risk of bias. The overall occurrence of any grade IVH and severe IVH among preterm infants was calculated along with a 95% CI. The temporal distribution of the onset of IVH was analyzed by pooling the time windows 0 to 6, 0 to 12, 0 to 24, 0 to 48, and 0 to 72 HOL. A subgroup analysis was conducted using studies published before and after 2007 to allow comparison with the results of a previous meta-analysis. A total of 21 567 records were identified, of which 64 studies and data from 9633 preterm infants were eligible. The overall rate of IVH did not decrease significantly before vs after 2007 (36%; 95% CI, 30%-42% vs 31%; 95% CI, 25%-36%), nor did severe IVH (10%; 95% CI, 7%-13% vs 11%; 95% CI, 8%-14%). The proportion of very early IVH (up to 6 HOL) after 2007 was 9% (95% CI, 3%-23%), which was 4 times lower than before 2007 (35%; 95% CI, 24%-48%). IVH up to 24 HOL before and after 2007 was 44% (95% CI, 31%-58%) and 25% (95% CI, 15%-39%) and up to 48 HOL was 82% (95% CI, 65%-92%) and 50% (95% CI, 34%-66%), respectively. This systematic review and meta-analysis found that the overall prevalence of IVH in preterm infants has not changed significantly since 2007, but studies after 2007 showed a later onset as compared with earlier studies, with only a small proportion of IVHs occurring before 6 HOL.

The Burr distribution as a model for the delay between key events in an individual's infection history.

Understanding the temporal relationship between key events in an individual's infection history is crucial for disease control. Delay data between events, such as infection and symptom onset times, is doubly censored because the exact time at which these key events occur is generally unknown. Current mathematical models for delay distributions are derived from heuristic justifications. Here, we derive a new model for delay distributions, specifically for incubation periods, motivated by bacterial-growth dynamics that lead to the Burr family of distributions being a valid modelling choice. We also incorporate methods within these models to account for the doubly censored data. Our approach provides biological justification in the derivation of our delay distribution model, the results of fitting to data highlighting the superiority of the Burr model compared to currently used models when the mode of the distribution is clearly defined or when the distribution tapers off. Under these conditions, our results indicate that the derived Burr distribution is a better-performing model for incubation-period data than currently used methods, with the derived Burr distribution being 13 times more likely to be a better-performing model than the gamma distribution for Legionnaires' disease based on data from a known outbreak.

THE DEADLINE FOR DEPARTURE PREVENTION FROM AN IMMIGRATION LAW PERSPECTIVE

This study aims to highlight the critical need to update the provisions on departure prevention in Indonesia's Immigration Law, specifically addressing the legal uncertainty arising from the phrase "at any time" in Article 97 Paragraph (1), which has the potential to cause prevention without an exact time limit. The study utilized a qualitative approach, employing a combination of legal analysis and empirical observation. The case approach allowed the author to analyze court decisions or concrete cases related to departure prevention, providing insights into how these regulations were implemented in practice and their implications for individual rights. The analysis focused on the implications of these legal frameworks for individuals' rights and the effectiveness of departure prevention measures, drawing connections between legal norms and their practical outcomes. The research provides a foundation for further discussions on enhancing legal certainty and protecting individual rights within the immigration framework, thereby contributing to the overall improvement of the legal system in Indonesia.

The provenance and provenience of photographs: Reconnecting serial images in the Umzimkhulu district, South Africa, 1899–2023

ABSTRACT This article introduces the ideas of provenance and provenience as a methodological approach to the study of photography from and in colonial contexts. It argues that rendering visible in this way the complex microhistories and photographic occasions of colonization, is an essential precondition to considering what a decolonization of photography could actually look like. Retracing a photograph’s trajectory back in time as image, object, and practice at times not only allows it to be reconnected to the exact place and moment in which it has been produced, but also shows its entanglement with other objects and actors as a lived experience. While this approach has value for studying photography at large, it is developed here through a series of 17 photographs showing a so-called first-fruits ceremony (ingcubhe), all taken on the same photographic occasion in the South African Umzimkhulu district in March 1899. Only by matching historical and present photographic frames and framings are descendants of the photographed and other stakeholders able to engage collectively with the photographs and their histories. It follows that it is more appropriate to the medium’s nature to think through the process of returning such photographs to their place of origin, not as a visual repatriation, but first of all as a visual reconnection. Reconnecting historical photographs to the exact space and time of their production through the research method repeat photography eventually creates new photographic occasions, which in turn are able to socially reanimate old ones.

The nature, sequence and duration of professional activities of Emergency Medical Service providers: An observational study to evaluate quality of care using the steps in the EMS care process as described by the SPART model.

The quality of care given by EMS staff is in part monitored by routine review of deployment reports that the providers must complete after each deployment. The assessment consists of determining whether a deployment was performed in a medically correct manner and thus whether the clinical reasoning process was appropriate. The time used during the deployment is also measured, as there are often time limits for performing different parts of the deployment and these must be adhered to. One might question whether measuring the time spent really gives insight into the quality of care, and if it might not be more useful to see how that time was spent. We therefore conceived a study to map the EMS care process and evaluate this process in relation to the time used. We used a focused ethnographic study design. Video recordings of EMS deployments in combination with semi-structured interviews of the EMS providers involved, were thematically analysed. This led to insights into the EMS care process and the embedded clinical reasoning. By categorising the professional activities we identified we were able to define a model that describes in general the EMS care process. We reported the first part of this study in which we developed this so called SPART model in a separate paper. In this second part of the study we determined the execution time within a deployment and measured the frequency of occurrence and the duration of the identified activities. We interpreted the operational content and the time variables both qualitatively and quantitatively. We also determined the distribution of activities over the three deployment operational periods (response, on-scene and transport period). Using the SPART model, we gained insight into the different activities' nature, order, and duration. We could qualitatively judge the effectiveness of the clinical reasoning process, i.e., the quality of care delivered. Generally, the studied cases were followable, and the clinical reasoning process was medically logical. The diagnostic process sometimes continued after the clinical decision, which was not medically logical. Remarkably, this never changed the clinical decision. Although this could negatively affect the quality of care, we found no clinical evidence that this was the case. Our findings demonstrated that the quality of care in EMS can be measured by using the SPART model to evaluate EMS deployments. We concluded that qualitative judgment was more important than quantitative evaluation. Interpreting the order of different activities led to the clearest understanding of the clinical reasoning process. It was concluded that knowledge of the exact time used per activity and, in total, had the least impact on understanding the clinical reasoning process.

FRAME – A FRAMEwork for Objectively Measuring Fear Based on Physiological and Psychological Data

Abstract Psychological trials, such as behavioural avoidance tests (BAT), are a fundamental part of the phobia therapeutic process. In order to link physiological reactions with specific points in time during psychological trials, it is necessary to integrate observation data with data collected automatically by sensors, such as wearable devices. To this end, this paper introduces FRAME - a framework for combining real-world events occurring during psychological trials with physiological data collected by wearables. FRAME consists of three parts, an Observation App, a data integration module and a Virtual Reality (VR) App. The Observation App captures events and their exact time of occurrence. The integration module links the observations with the respective physiological data, allowing an in-depth analysis of physiological reactions. The VR App provides a virtual scenario based on the BAT in vivo, thus enabling a BAT in virtuo. The practical applicability of FRAME is tested within a study comparing behavioural avoidance tests in vivo and in virtuo, assessing 25 patients with arachnophobia wearing an Empatica E4.

Chronological data dating the death of Khan Berke

The article considers the results of the source research of the texts of the Siro-Egyptian authors’ compiled writings. These texts were published in the fi rst volume of the Golden Horde’s compendium, translated in Russian by V.G. Tiesenhausen. Chronological evidence, contained in the writings’ messages were thoroughly researched using historical and chronological methods. Special attention is paid to applying the Cross-validation method to the chronological data, which allowed for the validation of the controversial evidence by comparing it with the information from external sources. The external sources in the research are the chronicles of Rashid-al-Din as well as Russian chronicles, which include the information about the death of Berke Khan. The results of the conducted research reconstruct the most exact time period of the death of Berke Khan. The revealed chronological controversies allowed to identify the more reliable and less accurate sources dating the events related to the period of Berke Khan reigning. Separate attention in the article is paid to the data on the kinship ties of the Horde khans. The author discusses the text compilations concerning the advantages and disadvantages connected with the details of the Siro-Egyptian authors’ compiled writings.

Chronological data dating the death of Khan Berke

The article considers the results of the source research of the texts of the Siro-Egyptian authors’ compiled writings. These texts were published in the fi rst volume of the Golden Horde’s compendium, translated in Russian by V.G. Tiesenhausen. Chronological evidence, contained in the writings’ messages were thoroughly researched using historical and chronological methods. Special attention is paid to applying the Cross-validation method to the chronological data, which allowed for the validation of the controversial evidence by comparing it with the information from external sources. The external sources in the research are the chronicles of Rashid-al-Din as well as Russian chronicles, which include the information about the death of Berke Khan. The results of the conducted research reconstruct the most exact time period of the death of Berke Khan. The revealed chronological controversies allowed to identify the more reliable and less accurate sources dating the events related to the period of Berke Khan reigning. Separate attention in the article is paid to the data on the kinship ties of the Horde khans. The author discusses the text compilations concerning the advantages and disadvantages connected with the details of the Siro-Egyptian authors’ compiled writings.

«Одним хорошим человеком стало меньше…»: отклики российской прессы на смерть академика А. Н. Пыпина

The work publishes the responses of St. Petersburg press to the death of a literary historian, academician Alexander Nikolaevich Pypin. The correspondence, notes, and obituaries of three newspapers – Peterburgskie Vedomosti, Rus’, and Novoye Vremya – for November 27–29, 1904 are presented. The materials are arranged in the chronological order of their publication in each edition. In newspaper correspondence, notes with exact date and time indication, event details, names of participants, a series of farewell events with A. N. Pypin were highlighted: the memorial service, funeral procession, burial. This reliable factual information fills the gap in the Pypin historiography related to the issues of his death, burial, as well as the first posthumous evaluations of his personality and activities expressed in the obituaries by his contemporaries. Periodicals express the unity of opinions about the late scientist, the social and historical significance of his work is noted. The obituaries emphasize the role of A. N. Pypin in educating the Russian society, in developing national consciousness, they highlight the popularity of his scientific research in various fields of scientific knowledge in wide learned circles, his international scientific authority. The name of the scientist serves as the personification of enlightenment, progress, genuine public spirit, and high humanistic ideals. The published newspaper materials make it possible to identify the attitude of the official authorities towards the scientist, to note their recognition of A. N. Pypin at the end of his life. The publication is accompanied by an introductory article and comments. The material is included in the collection of M. N. Chernyshevsky of the main fund of the Museum-Estate of N. G. Chernyshevsky (Saratov).

Solution of the space-fractional diffusion equationon bounded domains of superdiffusive phenomena.

Space-fractional diffusion equationsfind widespread application in nature. They govern the anomalous dynamics of many stochastic processes, generalizing the standard diffusion equationto superdiffusive behavior. Strikingly, the solution of space-fractional diffusion equationson bounded domains is still an open problem. This is in part due to the difficulty of handling nonlocal boundary conditions ascribed to the space-fractional derivative, leading to the failure of standard methods. Here we revisit the space-fractional diffusion equationin one spatial dimension with bounded domains and present a solution in terms of weighted Jacobi polynomials. Calculated eigenvalues and eigenfunctions in the superdiffusive regime show remarkable agreement with results from numerical discretization of the space-fractional derivative operator and Monte Carlo simulations. To exemplify, we apply the proposed solution to obtain the exact mean residence time or mean first-passage time, first-passage-time distribution, and survival probability, in agreement with known results for the superdiffusive regime. The system of equationsconverges rather fast for the first eigensolutions, as is desirable for practical application purposes in superdiffusive phenomena.

Inverse-Weighted Quantile Regression With Partially Interval-Censored Data.

This paper introduces a novel approach to estimating censored quantile regression using inverse probability of censoring weighted (IPCW) methodology, specifically tailored for data sets featuring partially interval-censored data. Such data sets, often encountered in HIV/AIDS and cancer biomedical research, may include doubly censored (DC) and partly interval-censored (PIC) endpoints. DC responses involve either left-censoring or right-censoring alongside some exact failure time observations, while PIC responses are subject to interval-censoring. Despite the existence of complex estimating techniques for interval-censored quantile regression, we propose a simple and intuitive IPCW-based method, easily implementable by assigning suitable inverse-probability weights to subjects with exact failure time observations. The resulting estimator exhibits asymptotic properties, such as uniform consistency and weak convergence, and we explore an augmented-IPCW (AIPCW) approach to enhance efficiency. In addition, our method can be adapted for multivariate partially interval-censored data. Simulation studies demonstrate the new procedure's strong finite-sample performance. We illustrate the practical application of our approach through an analysis of progression-free survival endpoints in a phase III clinical trial focusing on metastatic colorectalcancer.

Ms.FPOP: A Fast Exact Segmentation Algorithm with a Multiscale Penalty

Given a time series in R n with a piecewise constant mean and independent noises, we propose an exact dynamic programming algorithm for minimizing a least-squares criterion with a multiscale penalty, favoring well-spread changepoints. This penalty was proposed by Verzelen et al. and achieves optimal rates for changepoint detection and changepoint localization in a non-asymptotic scenario. Our proposed algorithm, Multiscale Functional Pruning Optimal Partitioning (Ms.FPOP), extends functional pruning ideas presented in Rigaill and Maidstone et al. to multiscale penalties. For large signals ( n ≥ 10 5 ) with sparse changepoints, Ms.FPOP is shown empirically to be quasi-linear and faster than the Pruned Exact Linear Time (PELT) method of Killick et al. applied to the multiscale penalty of Verzelen et al. which exhibits quadratic slowdown in these cases. We propose an efficient implementation of Ms.FPOP coded in C++ interfaced with R that can segment profiles of up to n = 10 6 in a matter of seconds. Our algorithm works for slightly more general multiscale penalties. In particular, it allows a minimum segment length to be imposed. Using simple simulations we then show that where profiles are sufficiently large ( n ≥ 10 4 ), Ms.FPOP using the multiscale penalty of Verzelen et al. is typically more powerful than optimizing a least-squares criterion with the BIC penalty of Yao, a criterion that was shown by Fearnhead and Rigaill to perfom well across a wide range of scenarios. Supplementary materials for this article are available online.

CHANGEPOINTS DETECTION OF PANDEMIC WAVE IN REAL-TIME: APPLICATIONS TO THE TRANSMISSION OF COVID-19

The COVID-19 pandemic has spread throughout the world. Most countries experienced the pandemic in multiple waves. The Richards model predicts when a pandemic will peak and end in a particular area. However, this model can only be used in the single-wave case. The research aims to identify a changepoint detection method capable of delineating pandemic wave boundaries, thus enabling the resolution of multiple wave cases using the Richards model. This article uses two methods to detect changepoints: the Pruned Exact Linear Time (PELT) and the interpolation method. PELT method determines the changepoint based on changes in the statistical properties of the sequence of observations which can be in the form of differences in the mean or variance of each set of observations. In contrast, the linear interpolation method determines the changepoint based on the slope of a data pattern. The two methods complement each other, where the interpolation method is used to determine whether the pandemic is still in a single wave or has multiple waves, followed by determining wave boundaries using the PELT method. Richards model parameter estimation is carried out after the wave boundaries are obtained, and initial data is taken from the last wave using the PELT method. The prediction results show the peak of the pandemic in a particular region and when it will end, which can be used to inform medium-term strategies for the government to overcome the ongoing pandemic. This information helps prevent a resurgence of infections, which would negatively affect the COVID-19 mortality rate and the area's economic situation.

Heart rate variability to predict 1-day atrial fibrillation episode onset using machine learning

Abstract Background Studies using AI assessment of 10-second 12-lead ECG suggest that it provides incremental identification of AF risk compared with clinical scores. However, it is not clear what preventive treatment should be given as the exact time of AF onset remains unknown. The use of machine learning (ML) with 2-lead Holter ECG allows the development of predictive models for paroxysmal AF within a short time window, potentially allowing an optimised pill-in-the-pocket (PITP)-like strategy. Aim To identify patients who are still in sinus rhythm but will develop an episode of AF in the next few hours using 24-hour Holter recordings and ML. Materials and methods We created a new database of 95871 Holter recordings that were manually analysed. We identified 1319 paroxysmal AF episodes from 872 patients. A total of 835 AF episodes from 506 recordings had more than 60 minutes of normal sinus rhythm (NSR) before AF onset and more than 10 minutes of AF after onset, and a total of 964 AF episodes had 30 minutes or more of NSR before AF onset. Patients were divided into five groups: all patients, patients younger than 60 years, 60-70 years, 70-80 years, and older than 80 years. A total of 365 recordings from 347 patients without rhythm abnormalities were identified and classified in our database, from which we took two ECG windows. We used a gradient-boosted decision tree model with heart rate variability (HRV) parameters as input. Results The most significant results were obtained from recordings with more than 5 minutes of AF with an AUROC of 0.919 (95% CI: 0.879-0.958) and an AUPRC of 0.919 (0.879-0.958) 15H08 ± 13H05 hours before the onset of AF (Figures 1 - 2). Using a threshold of 0.5, the accuracy was 84.5% (81.2-87.8), the sensitivity was 83.0% (79.5-86.4), the specificity was 86.6% (79.3-93.9), the PPV was 90.2% (85.5-94.9), the NPV was 78.4% (74.7-82.1) and the F1 score was 86.2% (83.5-89.0) for all patients. Conclusion These results suggest that the most important information for short-term prediction of AF onset is contained in HRV, allowing a preventive strategy. This could be exploited by wearables in mHealth allowing the use of a PITP-like preventive strategy to reduce the burden of AF. Prospective studies are needed to confirm the encouraging potential of these findings.AUROC of the XGBoost modelPrecision-recall curves