Non-stationary spatio-temporal point process modeling for high-resolution COVID-19 data

Most COVID-19 studies commonly report figures of the overall infection at a state- or county-level. This aggregation tends to miss out on fine details of virus propagation. In this paper, we analyze a high-resolution COVID-19 dataset in Cali, Colombia, that records the precise time and location of every confirmed case. We develop a non-stationary spatio-temporal point process equipped with a neural network-based kernel to capture the heterogeneous correlations among COVID-19 cases. The kernel is carefully crafted to enhance expressiveness while maintaining model interpretability. We also incorporate some exogenous influences imposed by city landmarks. Our approach outperforms the state-of-the-art in forecasting new COVID-19 cases with the capability to offer vital insights into the spatio-temporal interaction between individuals concerning the disease spread in a metropolis.

A novel PRFB decomposition for non-stationary time-series and image analysis

This work presents a novel perfect reconstruction filterbank decomposition (PRFBD) method for nonlinear and non-stationary time-series and image data representation and analysis. The Fourier decomposition method (FDM), an adaptive approach based on Fourier representation (FR), is shown to be a special case of the proposed PRFBD. In addition, adaptive Fourier–Gauss decomposition (FGD) based on FR and Gaussian filters, and adaptive Fourier–Butterworth decomposition (FBD) based on Butterworth filters are developed as the other special cases of the proposed PRFBD method. The proposed theory of PRFBD can decompose any signal (time-series, image, or other data) into a set of desired number of Fourier intrinsic band functions (FIBFs) that follow the amplitude-modulation and frequency-modulation (AM-FM) representations. A generic filterbank representation, where perfect reconstruction can be ensured for any given set of lowpass or highpass filters, is also presented. We performed an extensive analysis on both simulated and real-life data (COVID-19 pandemic, Earthquake and Gravitational waves) to demonstrate the efficacy of the proposed method. The resolution results in the time-frequency representation demonstrate that the proposed method is more promising than the state-of-the-art approaches. © 2023 Elsevier B.V.

Prewhitening and normalization help detect a strong cross-correlation between daily wastewater SARS-CoV-2 RNA abundance and COVID-19 cases in a community

Wastewater surveillance is a promising technology for real-time tracking and even early detection of COVID-19 infections in a community. Although correlation analysis between wastewater surveillance data and the daily clinical COVID-19 case numbers has been frequently conducted, the importance of stationarity of the time series data has not been well addressed. In this study, we demonstrated that strong yet spurious correlation could arise from non-stationary time series data in wastewater surveillance. Data prewhitening to remove trends by the first differences of values between two consecutive times helped to reveal distinct cross-correlation patterns between daily clinical case numbers and daily wastewater SARS-CoV-2 RNA abundance during a lockdown period in 2020 in Honolulu, Hawaii. Normalization of wastewater SARS-CoV-2 RNA concentration by the endogenous fecal viral markers in the same samples significantly improved the cross-correlation, and the best correlation was detected at a two-day lag of the daily clinical case numbers. The detection of a significant correlation between the daily wastewater SARS-CoV-2 RNA abundance and the clinical case numbers also suggests that disease burden fluctuation in the community should not be excluded as a contributor to the often observed weekly cyclic patterns of clinical cases. © 2023 The Royal Society of Chemistry.

Estimation of SIRM Model's Nonstationary Parameters of COVID-19 in Czech Rebuplic

This paper is a comprehensive study dedicated to practical solution of estimation problems in models of the spread of infectious diseases. Mentioned algorithms of parameters' estimation make possible to build mathematical models of the spread of infectious diseases based on observations. The results of analysis of the approach to mathematical modeling of the spread of infectious diseases are given in this paper, in particular simulation models and estimation methods in models of population dynamics. Computer simulation for analysis of COVID-19 pandemic in Czech Republic demonstrates efficiency o f t he mentioned algorithm. © 2022 IEEE.

An Incremental Learning-based Framework for Non-stationary Traffic Representations Clustering and Forecasting

To curb the growth of COVID-19, many rules, including a work-from-home policy, were issued in 2020. While these limits successfully prevented the virus's transmission, they completely altered original mobility patterns, resulting in considerable reductions in travel time and vehicle miles traveled. Under this non-stationary data stream, the US Department of Transportation struggled to anticipate future traffic conditions. Obviously, two essential challenges need to be addressed immediately: 1) it is challenging for transportation agencies to learn representative traffic patterns from the continually changing traffic circumstances. And 2) when and how should the forecasting model be updated to learn new patterns without forgetting previous tasks? We proposed an incremental learning-based framework for non-stationary data clustering and forecasting in transportation scenarios to tackle the issues mentioned above. It is a dual-module architecture that includes a Temporal Neighborhood Clustering module and an Incremental Learning module. The objective of the first component is to dynamically detect the optimal boundary for clustering statistically similar neighbors by enlarging both the in-group similarity and between-group dissimilarity. The second module applies the online-EWC approach, which is commonly used in image classification tasks but rarely in regression models, to learn new tasks and avoid catastrophic forgetting, which is a typical occurrence while training neural networks with multiple tasks. Experiments on the Greater Seattle Area employed loop detector data collected in 2020 yielded reliable prediction performance in both robustness and accuracy. The dual-module framework can generate promising results from pre-COVID-19 to post-COVID-19 time frames. This framework would aid government agencies and the general public in developing long-term policies and strategies for post-pandemic intelligent transportation systems. © 2022 IEEE.

Interpretable Operations Research for High-Stakes Decisions: Designing the Greek COVID-19 Testing System

In the summer of 2020, in collaboration with the Greek government, we designed and deployed Eva-the first national-scale, reinforcement learning system for targeted COVID-19 testing. In this paper, we detail the rationale for three major design/algorithmic elements: Eva's testing supply chain, estimating COVID-19 prevalence, and test allocation. Specifically, we describe the design of Eva's supply chain to collect and process thousands of biological samples per day with special emphasis on capacity procurement. Then, we propose a novel, empirical Bayes estimation strategy to estimate COVID-19 prevalence among various passenger types with limited data and showcase how these estimates were instrumental in making a variety of downstream decisions. Finally, we propose a novel, multiarmed bandit algorithm that dynamically allocates tests to arriving passengers in a nonstationary environment with delayed feedback and batched decisions. All our design and algorithmic choices emphasize the need for transparent reasoning to enable human-in-the-loop analytics. Such transparency was crucial to building trust and acceptance among policymakers and public health experts in a period of global crisis.

Dynamic Inventory Replenishment with Reinforcement Learning in Managing E-Fulfilment Centres

The demand of retail e-commerce has been rapidly growing due to the digitalization and the COVID-19 pandemic, and thus, the stress on e-fulfilment services continues to increase nowadays. To fulfil daily customers’ orders, effective inventory replenishment is of the essence in order to strike a balance between inventory management costs and service level. This paper describes an enhanced inventory replenishment approach by using reinforcement learning to deal with non-stationary and uncertain demand from customers. The proposed approach relaxes the assumption of stationary demand distribution considered in typical inventory models. Conventional policies derived from such models cannot guarantee optimal re-order quantities, when demand distribution is non-stationary over time. Consequently, reinforcement learning is adopted in the proposed approach to improve feasible solutions continuously in a dynamic business environment. In comparison to the conventional base stock policy, our proposed approach provides cost saving opportunities ranging from 28.5 to 41.3% in a simulated environment. It is found that the value of using data-driven solution approaches to deal with the practical inventory management problem is effective. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

Pseudo Steady-State Period in Non-Stationary Infinite-Server Queue with State Dependent Arrival Intensity

An infinite-server queueing model with state-dependent arrival process and exponential distribution of service time is analyzed. It is assumed that the difference between the value of the arrival rate and total service rate becomes positive starting from a certain value of the number of customers in the system. In this paper, time until reaching this value by the number of customers in the system is called the pseudo steady-state period (PSSP). Distribution of duration of PSSP, its raw moments and its simple approximation under a certain scaling of the number of customers in the system are analyzed. Novelty of the considered problem consists of an arbitrary dependence of the rate of customer arrival on the current number of customers in the system and analysis of time until reaching from below a certain level by the number of customers in the system. The relevant existing papers focus on the analysis of time interval since exceeding a certain level until the number of customers goes down to this level (congestion period). Our main contribution consists of the derivation of a simple approximation of the considered time distribution by the exponential distribution. Numerical examples are presented, which confirm good quality of the proposed approximation.

Spatiotemporal disparities in regional public risk perception of COVID-19 using Bayesian Spatiotemporally Varying Coefficients (STVC) series models across Chinese cities.

Regional public attention has been critical during the COVID-19 pandemic, impacting the effectiveness of sub-national non-pharmaceutical interventions. While studies have focused on public attention at the national level, sub-national public attention has not been well investigated. Understanding sub-national public attention can aid local governments in designing regional scientific guidelines, especially in large countries with substantial spatiotemporal disparities in the spread of infections. Here, we evaluated the online public attention to the COVID-19 pandemic using internet search data and developed a regional public risk perception index (PRPI) that depicts heterogeneous associations between local pandemic risk and public attention across 366 Chinese cities. We used the Bayesian Spatiotemporally Varying Coefficients (STVC) model, a full-map local regression for estimating spatiotemporal heterogeneous relationships of variables, and improved it to the Bayesian Spatiotemporally Interacting Varying Coefficients (STIVC) model to incorporate space-time interaction non-stationarity at spatial or temporal stratified scales. COVID-19 daily cases (median contribution 82.6%) was the most critical factor affecting public attention, followed by urban socioeconomic conditions (16.7%) and daily population mobility (0.7%). After adjusting national and provincial impacts, city-level influence factors accounted for 89.4% and 58.6% in spatiotemporal variations of public attention. Spatiotemporal disparities were substantial among cities and provinces, suggesting that observing national-level public dynamics alone was insufficient. Multi-period PRPI maps revealed clusters and outlier cities with potential public panic and low health literacy. Bayesian STVC series models are systematically proposed and provide a multi-level spatiotemporal heterogeneous analytical framework for understanding collective human responses to major public health emergencies and disasters.

Forecasting of a complex phenomenon using stochastic data-based techniques under non-conventional schemes: The SARS-CoV-2 virus spread case.

Epidemics are complex dynamical processes that are difficult to model. As revealed by the SARS-CoV-2 pandemic, the social behavior and policy decisions contribute to the rapidly changing behavior of the virus' spread during outbreaks and recessions. In practice, reliable forecasting estimations are needed, especially during early contagion stages when knowledge and data are insipient. When stochastic models are used to address the problem, it is necessary to consider new modeling strategies. Such strategies should aim to predict the different contagious phases and fast changes between recessions and outbreaks. At the same time, it is desirable to take advantage of existing modeling frameworks, knowledge and tools. In that line, we take Autoregressive models with exogenous variables (ARX) and Vector autoregressive (VAR) techniques as a basis. We then consider analogies with epidemic's differential equations to define the structure of the models. To predict recessions and outbreaks, the possibility of updating the model's parameters and stochastic structures is considered, providing non-stationarity properties and flexibility for accommodating the incoming data to the models. The Generalized-Random-Walk (GRW) and the State-Dependent-Parameter (SDP) techniques shape the parameters' variability. The stochastic structures are identified following the Akaike (AIC) criterion. The models use the daily rates of infected, death, and healed individuals, which are the most common and accurate data retrieved in the early stages. Additionally, different experiments aim to explore the individual and complementary role of these variables. The results show that although both the ARX-based and VAR-based techniques have good statistical accuracy for seven-day ahead predictions, some ARX models can anticipate outbreaks and recessions. We argue that short-time predictions for complex problems could be attained through stochastic models that mimic the fundamentals of dynamic equations, updating their parameters and structures according to incoming data.

Early Detection of COVID-19 Hotspots Using Spatio-Temporal Data

Recently, the Centers for Disease Control and Prevention (CDC) has worked with other federal agencies to identify counties with increasing coronavirus disease 2019 (COVID-19) incidence (hotspots) and offers support to local health departments to limit the spread of the disease. Understanding the spatio-temporal dynamics of hotspot events is of great importance to support policy decisions and prevent large-scale outbreaks. This paper presents a spatio-temporal Bayesian framework for early detection of COVID-19 hotspots (at the county level) in the United States. We assume both the observed number of cases and hotspots depend on a class of latent random variables, which encode the underlying spatio-temporal dynamics of the transmission of COVID-19. Such latent variables follow a zero-mean Gaussian process, whose covariance is specified by a non-stationary kernel function. The most salient feature of our kernel function is that deep neural networks are introduced to enhance the model's representative power while still enjoying the interpretability of the kernel. We derive a sparse model and fit the model using a variational learning strategy to circumvent the computational intractability for large data sets. Our model demonstrates better interpretability and superior hotspot-detection performance compared to other baseline methods. IEEE

The correlation analysis of the daily Covid-19 new cases data series in Albania

We analyzed herein the new covid-19 daily positive cases recorded in Albania. We observed that the distribution of the daily new cases is non-stationary and usually has a power law behavior in the low incidence zone, and a bell curve for the remaining part of the incidence interval. We qualified this finding as the indicator intensive dynamics and as proof that up now, the heard immunity has not been reached. By parallelizing the preferential attachment mechanisms responsible for a power law distribution in the social graphs elsewhere, we explain the low daily incidence distribution as result of the imprudent gatherings of peoples. Additionally, the bell-shaped distribution observed for the high daily new cases is agued as outcome of the competition between illness advances and restriction measures. The distribution is acceptably smooth, meaning that the management has been accommodated appropriately. This behavior is observed also for two neighbor countries Greece and Italy respectively, but was not observed for Turkey, Serbia, and North Macedonia. Next, we used the multifractal analysis to conclude about the features related with heterogeneity of the data. We have identified the local presence self-organization behavior in some separate time intervals. Formally and empirically we have identified that the full set of the data contain two regimes finalized already, followed by a third one which started in July 2021. © 2021 Institute of Physics Publishing. All rights reserved.

The modelling of heteroscedastics IDR-USD exchange rate with intervention and outlier factors

The nonstationary in time series data may be caused by the existence of intervention, outliers, and heteroscedastic effects. The outliers can represent an intervention so that it creates a heteroscedastic process. This research investigates the involvements of these three factors in time series data modelling. It is also reviewed how long the effects of the intervention and outliersfactors will last. The weekly IDR-USD exchange rate in period of May 2015 to April 2020 be evaluated. It is obtained that ARIMA model with the intervention factor gives the best re-estimation result, with smallest average of errors squared. Meanwhile for prediction, the heteroscedastic effect combined with outlier factors gives better results with the lowest percentage of errors. One of the phenomenal interventions in this data is the Covid-19 pandemic, which was started in Indonesia on March 2020. It is found that the effect of the intervention lasts less than five months and the prediction shows that the volatility of IDR-USD exchange rate starts to decline. This shows the stability of the process is starting to be maintained. © Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence.

REGULATORY POLICY TRANSFORMATION IN CONDITIONS OF NON-STATIONARY ECONOMY IN EASTERN EUROPEAN COUNTRIES: PRACTICAL APPROACH

European economic processes have always been the center of public interest. 2010-2019 were characterized by economic development and result achievements in the conditions of information flows exchange and digital changes. However, the pandemic effects of Covid-19 have negatively influenced economic activity, causing significant uncertainty in subsequent economic processes, which were reflected in GDP trends and determined the non-stationary economic conditions. And although the first half of 2021 marks GDP growth, the economic recovery, projected in the traditional scenario, is not enough to reach the precrisis level of production again. The purpose of the study is to identify the effects on European countries economic development and to develop proposals for directions and instruments of regulatory policy transformation, which would take into account the current non-stationary economic conditions. Achieving the goal led to the usage of scientific and practical methods of cognition, including the method of deduction and induction, system analysis, synthesis, generalization, mathematical methods and models. The study characterizes the nature of non-stationary economic development and identifies the need for regulatory influence to ensure further economic growth. The analysis proposes to use GDP as an indicator of economic processes dynamics with the specification of system of direct and indirect influence factors using a multifactor dynamic model. The assessment of depth and nature of their impact allowed to divide them by the stationary and non-stationary criterion. On the basis of received values the directions of regulatory policy transformation in the conditions of non-stationary economic processes have been offered.

Impact of COVID-19 lockdown on the particulate matter over Perungudi, Chennai, India.

The nationwide lockdown of Phase-1 in India was started from March 25 to April 14, 2020 and Phase-2 from April 15 to May 3, 2020 with severe restrictions on public activities in India. Utilizing the particulate matter PM10 and PM2.5 data recorded during this adverse time, the present study is undertaken to assess the impact of phase 1 and 2 lockdown on the air quality of Perungudi, Chennai, India. The data obtained from the Tamil Nadu Pollution Control Board was assessed for lockdown phase. We compared particulate matter data for the unlock phase with a coinciding period in March 2020 to determine the changes in pollutant concentrations during the lockdown period of April 2020. The descriptive analysis of PM continuous data was performed to determine the mean, standard deviation, variance, skew and kurtosis to identify the nature of data. Correlogram analysis gives the information that the data under study has non-stationary behaviour and not random. Along with this linear regression analysis were performed to determine the relationship and trend for the data. The results revealed decreasing trend in the concentrations (PM10, PM2.5).

A novel analysis of COVID 19 risk in India incorporating climatic and socioeconomic Factors.

This study investigates the influence of climate variables (pressure, relative humidity, temperature and wind speed) in inducing risk due to COVID 19 at rural, urban and total (rural and urban) population scale in 623 pandemic affected districts of India incorporating the socioeconomic vulnerability factors. We employed nonstationary extreme value analysis to model the different quantiles of cumulative COVID 19 cases in the districts by using climatic factors as covariates. Wind speed was the most dominating climatic factor followed by relative humidity, pressure, and temperature in the evolution of the cases. The results reveal that stationarity, i.e., the COVID 19 cases which are independent of pressure, relative humidity, temperature and wind speed, existed only in 148 (23.7%) out of 623 districts. Whereas, strong nonstationarity, i.e., climate dependence, was detected in the cases of 474 (76.08%) districts. 334 (53.6%), 200 (32.1%) and 336 (53.9%) districts out of 623 districts were at high risk (or above) at rural, urban and total population scales respectively. 19 out of 35 states were observed to be under high (or above) Kerala, Maharashtra, Goa and Delhi being the most risked ones. The study provides high-risk maps of COVID 19 pandemic at the district level and is aimed at supporting the decision-makers to identify climatic and socioeconomic factors in augmenting the risks.