Concept Drift Adaptation Methods under the Deep Learning Framework: A Literature Review

With the advent of the fourth industrial revolution, data-driven decision making has also become an integral part of decision making. At the same time, deep learning is one of the core technologies of the fourth industrial revolution that have become vital in decision making. However, in the era of epidemics and big data, the volume of data has increased dramatically while the sources have become progressively more complex, making data distribution highly susceptible to change. These situations can easily lead to concept drift, which directly affects the effectiveness of prediction models. How to cope with such complex situations and make timely and accurate decisions from multiple perspectives is a challenging research issue. To address this challenge, we summarize concept drift adaptation methods under the deep learning framework, which is beneficial to help decision makers make better decisions and analyze the causes of concept drift. First, we provide an overall introduction to concept drift, including the definition, causes, types, and process of concept drift adaptation methods under the deep learning framework. Second, we summarize concept drift adaptation methods in terms of discriminative learning, generative learning, hybrid learning, and others. For each aspect, we elaborate on the update modes, detection modes, and adaptation drift types of concept drift adaptation methods. In addition, we briefly describe the characteristics and application fields of deep learning algorithms using concept drift adaptation methods. Finally, we summarize common datasets and evaluation metrics and present future directions.

A data-driven model for sustainable and resilient supplier selection and order allocation problem in a responsive supply chain: A case study of healthcare system

This research attempts to study the Supplier Selection and Order Allocation Problem (SSOAP) considering three crucial concepts, namely responsiveness, sustainability, and resilience. To do so, the current research develops a Multi-Stage Decision-Making Framework (MSDMF) to select potential suppliers and determine the quantity of orders. The first stage aims at computing the scores of the suppliers based on several indicators. To do this, a novel decision-making approach named the Stochastic Fuzzy Best–Worst Method (SFBWM) is developed. Then, in the second stage, a Multi-Objective Model (MOM) is suggested to deal with supplier selection and order allocation decisions. In the next step, a data-driven Fuzzy Robust Stochastic (FRS) optimization approach, based on the fuzzy robust stochastic method and the Seasonal Autoregressive Integrated Moving Average (SARIMA) methods, is employed to efficiently treat the hybrid uncertainty of the problem. Afterwards, a novel solution method named the developed Chebyshev Multi-Choice Goal Programming with Utility Function (CMCGP-UF) is developed to obtain the optimal solution. Moreover, given the crucial role of the Medical Equipment (ME) industry in society's health, especially during the recent Coronavirus disease, this important industry is taken into account. The outcomes of the first stage demonstrate that agility, cost, GHG emission, quality, robustness, and Waste Management (WM), respectively, are the most important criteria. The outcomes of the second stage determine the selected suppliers, utilized transportation systems, and established sites. It is also revealed that demand directly affects all the objective functions while increasing the rate of disruptions has a negative effect on the sustainability measures.

Misleading Beyond Visual Tricks: How People Actually Lie with Charts

Data visualizations can empower an audience to make informed decisions. At the same time, deceptive representations of data can lead to inaccurate interpretations while still providing an illusion of data-driven insights. Existing research on misleading visualizations primarily focuses on examples of charts and techniques previously reported to be deceptive. These approaches do not necessarily describe how charts mislead the general population in practice. We instead present an analysis of data visualizations found in a real-world discourse of a significant global event - Twitter posts with visualizations related to the COVID-19 pandemic. Our work shows that, contrary to conventional wisdom, violations of visualization design guidelines are not the dominant way people mislead with charts. Specifically, they do not disproportionately lead to reasoning errors in posters' arguments. Through a series of examples, we present common reasoning errors and discuss how even faithfully plotted data visualizations can be used to support misinformation. © 2023 Owner/Author.

Maintainers of Stability: The Labor of China's Data-Driven Governance and Dynamic Zero-COVID

This paper examines the social, technological, and emotional labor of maintaining China's data-driven governance broadly, and dynamic zero-COVID management in particular. Drawing on ethnographic research in China, we examine the sociotechnical work of maintenance during the 2022 Shanghai lockdown. This labor included coordinating mass testing, quarantine, and lockdown procedures as well as implementing ad-hoc technological workarounds and managing public sentiments. We demonstrate that, far from being effected from the top down, China's data-driven governance relies on the circumscribed participation of citizens. During Shanghai's lockdown, citizens with relevant expertise helped to maintain technological stability by fixing or programming data systems, but also to ensure the ongoing production of"positive feelings"about social stability through data-driven governance. In so doing, such citizens simultaneously enacted an ambivalent and circumscribed form of agency, and maintained social and by extension political stability. This article sheds light on data-driven governance and political processes of maintenance. © 2023 ACM.

Flattening the Curve after the Initial Outbreak of Coronavirus Disease 2019: A Data-Driven Modeling Analysis for the Omicron Pandemic in China.

China is relaxing COVID-19 measures from the "dynamic zero tolerance" (DZT) level. The "flatten-the-curve" (FTC) strategy, which decreases and maintains the low rate of infection to avoid overwhelming the healthcare system by adopting relaxed nonpharmaceutical interventions (NPIs) after the outbreak, has been perceived as the most appropriate and effective method in preventing the spread of the Omicron variant. Hence, we established an improved data-driven model of Omicron transmission based on the age-structured stochastic compartmental susceptible-latent-infectious-removed-susceptible model constructed by Cai to deduce the overall prevention effect throughout China. At the current level of immunity without the application of any NPIs, more than 1.27 billion (including asymptomatic individuals) were infected within 90 days. Moreover, the Omicron outbreak would result in 1.49 million deaths within 180 days. The application of FTC could decrease the number of deaths by 36.91% within 360 days. The strict implementation of FTC policy combined with completed vaccination and drug use, which only resulted in 0.19 million deaths in an age-stratified model, will help end the pandemic within about 240 days. The pandemic would be successfully controlled within a shorter period of time without a high fatality rate; therefore, the FTC policy could be strictly implemented through enhancement of immunity and drug use.

A data-driven optimization model to response to COVID-19 pandemic: a case study.

COVID-19 is a highly prevalent disease that has led to numerous predicaments for healthcare systems worldwide. Owing to the significant influx of patients and limited resources of health services, there have been several limitations associated with patients' hospitalization. These limitations can cause an increment in the COVID-19-related mortality due to the lack of appropriate medical services. They can also elevate the risk of infection in the rest of the population. The present study aims to investigate a two-phase approach to designing a supply chain network for hospitalizing patients in the existing and temporary hospitals, efficiently distributing medications and medical items needed by patients, and managing the waste created in hospitals. Since the number of future patients is uncertain, in the first phase, trained Artificial Neural Networks with historical data forecast the number of patients in future periods and generate scenarios. Through the use of the K-Means method, these scenarios are reduced. In the second phase, a multi-objective, multi-period, data-driven two-stage stochastic programming is developed using the acquired scenarios in the previous phase concerning the uncertainty and disruption in facilities. The objectives of the proposed model include maximizing the minimum allocation-to-demand ratio, minimizing the total risk of disease spread, and minimizing the total transportation time. Furthermore, a real case study is investigated in Tehran, the capital of Iran. The results showed that the areas with the highest population density and no facilities near them have been selected for the location of temporary facilities. Among temporary facilities, temporary hospitals can allocate up to 2.6% of the total demand, which puts pressure on the existing hospitals to be removed. Furthermore, the results indicated that the allocation-to-demand ratio can remain at an ideal level when disruptions occur by considering temporary facilities. Our analyses focus on: (1) Examining demand forecasting error and generated scenarios in the first phase, (2) exploring the impact of demand parameters on the allocation-to-demand ratio, total time and total risk, (3) investigating the strategy of utilizing temporary hospitals to address sudden changes in demand, (4) evaluating the effect of disruption to facilities on the supply chain network.

Development of a data-driven digital phenotype profile of distress experience of healthcare workers during COVID-19 pandemic.

BACKGROUND AND OBJECTIVE: Due to the constraints of the COVID-19 pandemic, healthcare workers have reported acting in ways that are contrary to their moral values, and this may result in moral distress. This paper proposes the novel digital phenotype profile (DPP) tool, developed specifically to evaluate stress experiences within participants. The DPP tool was evaluated using the COVID-19 VR Healthcare Simulation of Stress Experience (HSSE) dataset (NCT05001542), which is composed of passive physiological signals and active mental health questionnaires. The DPP tool focuses on correlating electrocardiogram, respiration, photoplethysmography, and galvanic skin response with moral injury outcome scale (Brief MIOS). METHODS: Data-driven techniques are encompassed to develop a tool for robust evaluation of distress among participants. To accomplish this, we applied pre-processing techniques which involved normalization, data sanitation, segmentation, and windowing. During feature analysis, we extracted domain-specific features, followed by feature selection techniques to rank the importance of the feature set. Prior to classification, we employed k-means clustering to group the Brief MIOS scores to low, moderate, and high moral distress as the Brief MIOS lacks established severity cut-off scores. Support vector machine and decision tree models were used to create machine learning models to predict moral distress severities. RESULTS: Weighted support vector machine with leave-one-subject-out-cross-validation evaluated the separation of the Brief MIOS scores and achieved an average accuracy, precision, sensitivity, and F1 of 98.67%, 98.83%, 99.44%, and 99.13%, respectively. Various machine learning ablation tests were performed to support our results and further enhance the understanding of the predictive model. CONCLUSION: Our findings demonstrate the feasibility to develop a DPP tool to predict distress experiences using a combination of mental health questionnaires and passive signals. The DPP tool is the first of its kind developed from the analysis of the HSSE dataset. Additional validation is needed for the DPP tool through replication in larger sample sizes.

Intellectual humility predicts COVID-19 preventive practices through greater adoption of data-driven information and feelings of responsibility

Preventive health practices have been crucial to mitigating viral spread during the COVID-19 pandemic. In two studies, we examined whether intellectual humility-openness to one's existing knowledge being inaccurate-related to greater engagement in preventive health practices (social distancing, handwashing, mask-wearing). In Study 1, we found that intellectually humble people were more likely to engage in COVID-19 preventive practices. Additionally, this link was driven by intellectually humble people's tendency to adopt information from data-driven sources (e.g., medical experts) and greater feelings of responsibility over the outcomes of COVID-19. In Study 2, we found support for these relationships over time (2 weeks). Additionally, Study 2 showed that the link between intellectual humility and preventive practices was driven by a greater tendency to adopt data-driven information when encountering it, rather than actively seeking out such information. These findings reveal the promising role of intellectual humility in making well-informed decisions during public health crises.

Mitigating the COVID-19 pandemic through data-driven resource sharing

COVID-19 outbreaks in local communities can result in a drastic surge in demand for scarce resources such as mechanical ventilators. To deal with such demand surges, many hospitals (1) purchased large quantities of mechanical ventilators, and (2) canceled/postponed elective procedures to preserve care capacity for COVID-19 patients. These measures resulted in a substantial financial burden to the hospitals and poor outcomes for non-COVID-19 patients. Given that COVID-19 transmits at different rates across various regions, there is an opportunity to share portable healthcare resources to mitigate capacity shortages triggered by local outbreaks with fewer total resources. This paper develops a novel data-driven adaptive robust simulation-based optimization (DARSO) methodology for optimal allocation and relocation of mechanical ventilators over different states and regions. Our main methodological contributions lie in a new policy-guided approach and an efficient algorithmic framework that mitigates critical limitations of current robust and stochastic models and make resource-sharing decisions implementable in real-time. In collaboration with epidemiologists and infectious disease doctors, we give proof of concept for the DARSO methodology through a case study of sharing ventilators among regions in Ohio and Michigan. The results suggest that our optimal policy could satisfy ventilator demand during the first pandemic's peak in Ohio and Michigan with 14% (limited sharing) to 63% (full sharing) fewer ventilators compared to a no sharing strategy (status quo), thereby allowing hospitals to preserve more elective procedures. Furthermore, we demonstrate that sharing unused ventilators (rather than purchasing new machines) can result in 5% (limited sharing) to 44% (full sharing) lower expenditure, compared to no sharing, considering the transshipment and new ventilator costs.

A Contact Matrix-Based Approach for Predicting COVID-19 Using Influenza Data

The global pandemic Corona Virus Disease 2019 (COVID-19) has become one of the deadliest epidemics in human history, bringing enormous harm to human society. To help health policymakers respond to the threat of COVID-19, prediction of outbreaks is needed. Research on COVID-19 prediction usually uses data-driven models and mechanism models. However, in the early stages of the epidemic, there were not enough data to establish a data-driven model. The inadequate understanding of the virus that causes COVID-19, SARS-COV-2, has also led to the inaccuracies of the mechanism model. This has left the government with the toughest Non-pharmaceutical interventions (NPIs) to curb the spread of the virus, such as the lockdown of Wuhan in 2020. Yet man is a social animal, and social relations and interactions are necessary for his existence. The novel coronavirus and containment measures have challenged human and community interactions, affecting the lives of individuals and collective societies. To help governments take appropriate and necessary actions in the early stages of an epidemic, and to mitigate its impact on people's psychology and lives, we used the COVID-19 pandemic as an example to develop a model that uses surveillance data from one epidemic to predict the development trend of another. Based on the fact that both influenza and COVID-19 are transmitted through infectious respiratory droplets, we hypothesized that they may have the same underlying contact structure, and we proposed the influenza data-based COVID-19 prediction (ICP) model. In this model, the underlying contact pattern is firstly inferred by using a singular value decomposition method from influenza surveillance data. Then the contact matrix was used to simulate the influenza virus transmission through close contact of people, and the influenza virus transmission model was established. In order to be able to simulate the spread of COVID-19 virus using influenza transmission models, we used influenza contact matrix and COVID-19 infection data to estimate the risk of a population contracting COVID-19, i.e. force of infection of COVID-19. Finally, we used force of infection and influenza virus transmission model to simulate and predict the spread of COVID-19 in the population. We obtained age-disaggregated influenza and COVID-19 infection data for the United States in 2020, as well as data for Europe, which was not disaggregated by age. We use correlation coefficients as an evaluation indicator, and the final results prove that the predicted value and the actual value are positively correlated. So, the development trend of COVID-19 can be predicted using influenza surveillance data. © 2022 IEEE.

Building a collaborative cloud platform to accelerate heart, lung, blood, and sleep research.

Research increasingly relies on interrogating large-scale data resources. The NIH National Heart, Lung, and Blood Institute developed the NHLBI BioData CatalystⓇ (BDC), a community-driven ecosystem where researchers, including bench and clinical scientists, statisticians, and algorithm developers, find, access, share, store, and compute on large-scale datasets. This ecosystem provides secure, cloud-based workspaces, user authentication and authorization, search, tools and workflows, applications, and new innovative features to address community needs, including exploratory data analysis, genomic and imaging tools, tools for reproducibility, and improved interoperability with other NIH data science platforms. BDC offers straightforward access to large-scale datasets and computational resources that support precision medicine for heart, lung, blood, and sleep conditions, leveraging separately developed and managed platforms to maximize flexibility based on researcher needs, expertise, and backgrounds. Through the NHLBI BioData Catalyst Fellows Program, BDC facilitates scientific discoveries and technological advances. BDC also facilitated accelerated research on the coronavirus disease-2019 (COVID-19) pandemic.

Population differences in completeness and reliability of Dutch COVID-19 registry data: The influence of testing policy, test readiness, and registration policy from an epidemiological perspective.

During the COVID-19 pandemic, the bidirectional relationship between policy and data reliability has been a challenge for researchers of the local municipal health services. Policy decisions on population specific test locations and selective registration of negative test results led to population differences in data quality. This hampered the calculation of reliable population specific infection rates needed to develop proper data driven public health policy.

Multi-Source Pandemic Data Visualization and Synchronization for Information Extraction

Visualization is integral to investigating information hidden in data and providing users with intuitive feedback for decision-making. No matter the field a data set describes, inspecting the data visually will yield fruitful insights into the trends and statistics. Over the past calendar year, COVID-19 vaccines have become increasingly available for much of the population. However, the CDC (Centers for Disease Control and Prevention) fails to consider multiple sets of pandemic data in a side-by-side view and synchronize multiple key factors in one web page, limiting medical professionals and individuals to seeing, comparing, and interacting with complete data visualization. To analyze the coronavirus and vaccination data collected from multiple sources, effectively displaying them is critically important for interpreting the pandemic transmission pattern and vaccine efficiency. This paper presents new algorithms for innovative data visualizations that provide users with intuitive feedback and enable them to see a complete story of where the data is concerned. The information derived from our developed web-based data visualization will aid healthcare professionals and everyday citizens in moving forward as the pandemic progresses. © 2023 IEEE.

A model for the SARS-CoV-2 dynamics in a population lacking herd immunity

We introduced the S-HI model, a generalized SEIR model to describe the dynamics of the SARS-CoV-2 virus in a community without herd immunity and performed simulations for six months. The S- HI model consists of eight equations corresponding to susceptible individuals, exposed, asymptomatic infected, asymptomatic recovered, symptomatic infected, quarantined, symptomatic recovered and dead. We study the dynamics of the infected, asymptomatic. Dead classes in 4 different networks: households, workplaces, agglomeration places and the general community, showing that the dynamics of the three compartments have the exact nature in each layer and that the speed of the disease considerably increases in the networks with the highest weight of contacts. The reproduction number, R0, is greater than 1 in all networks conforming to the theory. The variants of the SARS-Cov-2 virus are not taken into account, so the S-HI model would fit a situation similar to the first wave of contagion after the mandatory lockdown. Copyright: © 2022 by the authors.

Data-Driven Analysis and Optimization for Urban Energy Systems Equitable Resilience

Electric vehicles (EVs) can be leveraged as power resources to support the grid operation in challenging scenarios, e.g., natural disasters or health crises such as the COVID-19 pandemic. This paper aims to enhance equity of power resilience in urban energy systems by means of strategic allocation of EV charging infrastructure. We first use data-driven approaches to infer the relationships between communities' power resilience equity and available EV charging infrastructure as well as other prominent social-demographic factors. This inference leads to the development of a machine learning model for power resilience inequity prediction. We further develop an optimization frame-work that jointly considers equitable resiliency and resource utilization to guide the optimized EV charging infrastructure allocation across the city. Case studies demonstrate the capability of the devised approach in enhancing power resilience equity in marginalized communities. © 2023 IEEE.

The Role of Data-Driven Solutions for SMEs in Responding to COVID-19

This study aims to explore the role of small and medium-sized enterprises (SMEs) in developing data-driven solutions to address the direct and indirect challenges posed by COVID-19. A sample of six case studies of SMEs from the UK and Portugal were selected to explore in-depth the experience of these companies in proposing innovative solutions in the pandemic context. The findings reveal that the pandemic caused amplifying effects on the digitalization of organizations and the emergence of data-driven solutions. However, the development of a data-driven approach involves not only technologies but also the digitalization of processes and highly skilled human resources. The pandemic was also a catalyst for the emergence of collaborative initiatives that have enabled the development of solutions involving diverse players from science, business, and civilian society. This study offers innovative contributions by focusing exclusively on companies developing data-driven solutions supported by technologies such as the internet of things (IoT), big data, and artificial intelligence.

COVID-19 phenomics

Phenomic studies of coronavirus disease 2019 (COVID-19) attempt to comprehensively describe the range of phenotypes associated with disease-related outcomes, by either breadth or depth of characterization. The primary aims of such studies are the unbiased generation of hypotheses concerning COVID-19 pathophysiology and the empirical determination of effective prognostic indicators. Of particular relevance to COVID-19 are phenome-wide association studies—large-scale, data-driven studies evaluating associations between a multitude of phenotypic traits and COVID-19 severity or other outcomes of interest, often employing bioinformatic and statistical approaches for the analysis of databases of electronic health records. This type of extensive phenotyping, in combination with intensive interrogation of particular aspects of the pathophysiological response, also allows investigators to reconstruct the network of phenomena that underpin disease, of particular significance because of the systemic nature of COVID-19. Because of their ability to detect novel associations, another great utility of extensive phenomic analyses applied to COVID-19 is in the development of prognostic tools and biomarkers that improve the efficacy of patient care. Finally, when applied to those in the convalescent phase, phenomics has helped to elucidate both the nature of postacute sequalae of COVID-19 and the characteristics that predispose an individual toward them. Hence, phenomics provides an additional and unique perspective which is crucial to our understanding of COVID-19 to better equip us against unforeseen adverse outcomes of this pandemic and potential infectious outbreaks in the future. © 2023 Elsevier Inc. All rights reserved.

Preprocessing and Artificial Intelligence for Increasing Explainability in Mental Health

This paper shows the added value of using the existing specific domain knowledge to generate new derivated variables to complement a target dataset and the benefits of including these new variables into further data analysis methods. The main contribution of the paper is to propose a methodology to generate these new variables as a part of preprocessing, under a double approach: creating 2nd generation knowledge-driven variables, catching the experts criteria used for reasoning on the field or 3rd generation data-driven indicators, these created by clustering original variables. And Data Mining and Artificial Intelligence techniques like Clustering or Traffic light Panels help to obtain successful results. Some results of the project INSESS-COVID19 are presented, basic descriptive analysis gives simple results that even though they are useful to support basic policy-making, especially in health, a much richer global perspective is acquired after including derivated variables. When 2nd generation variables are available and can be introduced in the method for creating 3rd generation data, added value is obtained from both basic analysis and building new data-driven indicators. © 2023 World Scientific Publishing Company.

Data-driven Analysis of Remote Work in China during the COVID-19 Pandemic

This paper leverages online content to investigate the biggest impact of COVID-19 - remote work, by using China as a primary case study. Telecommuting has become popular since February 2020 primarily due to the pandemic, and people have been slowly returning to their office from May 2020. This study focuses on two time windows in the year 2020 to calculate the growth of different job sectors. Our results indicate the negative impact of teleworking in manufacturing industry, but shows that information technology-related industries are less affected by working from home. This paper also investigates the impact of COVID-19 on the stock market and discussed what plan of action the policy makers should take to provide a good economic environment for the country. In addition to the overall economic situation, we observed how the psychological situation of employees could affect their job performance, indirectly affecting the development of certain industry sectors. Therefore, misinformation in certain Chinese social media channels was also studied in this paper specifically examining the rumors and their latent topics. We believe that our work will initiate a dialogue between scientists, policy makers and government officials to consider the observations highlighted in this paper. © 2022 IEEE Computer Society. All rights reserved.

Systems biology in COVID-19

Coronavirus disease 2019 (COVID-19), the disease cause by the novel severe acute respiratory syndrome coronavirus 2 represents a global, unresolved challenge for researchers and clinicians alike. In the shadow of overwhelmed healthcare systems, the pressure to produce knowledge, standard operating procedures, efficacious treatments, and prophylactic agents has been unlike any other occasion in recent history. Systems biology, an assortment of methods that aim to model biological systems and their properties has risen to meet this multifaceted challenge. In this chapter, we review approaches and breakthroughs of systems biology research in COVID-19, along with the nascent clade of phenomics, a deep-phenotyping systems concept that has enabled the real-time integration of big data and analytical methods in clinical decision making. © 2023 Elsevier Inc. All rights reserved.