Risk Modeling Research Articles

Risk Modeling of Time-Varying Covariates Using an Ensemble of Survival Trees: Predicting Future Cancer Events

The challenge of survival prediction is ubiquitous in medicine, but only a handful of methods are available for survival prediction based on time-varying data. Here we propose a novel method for this problem, using a random forest of survival trees for left-truncated and right-censored data. We demonstrate the advantage of our method on prediction of breast cancer and prostate gland cancer risk among healthy individuals by analyzing routine laboratory measurements, vital signs and age. We analyze electronic medical records of 20,317 healthy individuals who underwent routine checkups and identified those who later developed cancer. In cross-validation, our method predicted future prostate and breast cancers six months before diagnosis with an area under the ROC curve of 0.62±0.05 and 0.6±0.03 respectively, outperforming standard random forest, random survival forest, cox-regression model, dynamic deep-hit and a single survival tree. Our work proposes a new framework for survival risk prediction in time-varying data and our results suggest that computational analysis of data on healthy individuals can improve the detection of those at risk of future cancer development.

Propensity score matching: a tool for consumer risk modeling and portfolio underwriting

Researchers and practitioners in financial services utilize a wide range of empirical techniques to assess risk and value. In cases where known performance is used to predict future performance of a new asset, the risk of bias is present when samples are uncontrolled by the analyst. Propensity score matching is a statistical methodology commonly used in medical and social science research to address issues related to experimental design when random assignment of cases is not possible. This common method has been almost absent from financial risk modeling and portfolio underwriting, primarily due to the different objectives for this sector relative to medicine and social sciences. In this application note, we demonstrate how propensity score matching can be considered as a practical tool to inform portfolio underwriting outside of experimental design. Using a portfolio of distressed consumer credit accounts, we demonstrate that propensity score matching can be used to predict both account-level and portfolio-level risk and argue that propensity score matching should be included in the methodological toolbox of researchers and practitioners engaged in risk modeling and valuation activities of portfolios of consumer assets, particularly in contexts with limited observations, a large number of potential modeling features, or highly imbalanced covariates.

Structured dictionary learning of rating migration matrices for credit risk modeling

Structured dictionary learning of rating migration matrices for credit risk modeling

Competing Risk Modeling with Bivariate Varying Coefficients to Understand the Dynamic Impact of COVID-19

The coronavirus disease 2019 (COVID-19) pandemic has exerted a profound impact on patients with end-stage renal disease relying on kidney dialysis to sustain their lives. A preliminary analysis of dialysis patient postdischarge hospital readmissions and deaths in 2020 revealed that the COVID-19 effect has varied significantly with postdischarge time and time since the pandemic onset. However, the complex dynamics cannot be characterized by existing varying coefficient models. To address this issue, we propose a bivariate varying coefficient model for competing risks, where tensor-product B-splines are used to estimate the surface of the COVID-19 effect. An efficient proximal Newton algorithm is developed to facilitate the fitting of the new model to the massive data for Medicare beneficiaries on dialysis. Difference-based anisotropic penalization is introduced to mitigate model overfitting and effect wiggliness; a cross-validation method is derived to determine optimal tuning parameters. Hypothesis testing procedures are designed to examine whether the COVID-19 effect varies significantly with postdischarge time and the time since the pandemic onset, either jointly or separately. Applications to Medicare dialysis patients demonstrate the real-world performance of the proposed methods. Simulation experiments are conducted to evaluate the estimation accuracy, type I error rate, statistical power, and model selection procedures. Supplementary materials for this article are available online.

Modeling and management of cyber risk: a cross-disciplinary review

Abstract This paper provides a review of cyber risk research accomplished in different disciplines, with a primary goal to aid researchers in the field of insurance and actuarial science in identifying potential research gaps as well as leveraging useful models and techniques that have been considered in the literature. We highlight the recent advancements in cyber risk prediction, modeling, management, and insurance achieved in different domains including computer engineering, actuarial science, and business studies. The surveyed works are classified according to their respective modeling approaches, allowing readers to more easily compare the technical aspects of the surveyed works and spot out research gaps based on the research tools of their liking. We conclude this paper with a summary of possible research directions that are identified from the review.

Ethical AI and the Future of Healthcare: Combining Academic Theory and Industry Practice to Ensure Patient-Centered Care

Purpose: This paper seeks to define a risk taxonomy, establish meaningful controls, and create a prospective harms model for AI risks in healthcare. Currently, there is no known comprehensive definition of AI risks, as applied to industry and society. 
 Materials and Methods: The temptation for current research, both in academia and industry, is to apply exclusively-tech-based solutions to these complex problems; however, this view is myopic, and can be remedied by establishing effective controls informed by a holistic approach to managing AI risk. Sociotechnical Systems Theory (STS) is an attractive theoretical lens for this issue, because it prevents collapsing a multifaceted problem into a one-dimensional solution. Specifically, the multidisciplinary approach—one that includes both the sciences and the humanities—reveals a multidimensional view of technology-society interaction, exemplified by the advent of AI. 
 Findings: After advancing this risk taxonomy, this paper utilizes the risk management framework of Lean Six Sigma (LSS) to propose effective mitigating controls for the identified risks. LSS determines controls through data collection and analysis, and supports data-driven decision making for industry professionals.
 Implications to Theory, Practice and Policy: Instantiating the theory of STS into industry practices could be critical, then, for determining and mitigating real-world risks from AI. In summary, this paper combines the academic theory of sociotechnical systems with the industry practice of Lean Six Sigma to develop a hybrid model to fill a gap in the literature. Drawing upon both theory and practice ensures a robust, informed risk model of AI use in healthcare.

Association between triglyceride glucose index and all-cause mortality in patients with cerebrovascular disease: a retrospective study

BackgroundTriglyceride glucose (TyG) is associated with stroke, atherosclerosis, and adverse clinical outcomes. However, its correlation with cerebrovascular disease (CVD) mortality remains unclear. This study aimed to investigate the relationship between TyG index and mortality in patients with CVD.MethodsPatient data sourced from the Medical Information Mart for Intensive Care -IV database were categorized based on TyG quartiles. Kaplan–Meier survival analysis was used to estimate survival disparities among the TyG subgroups. Cox proportional risk modeling was used to examine the association between the TyG index and mortality. Generalized summation models were applied to fit the smoothed curves. log-likelihood ratio test were used to analyze the non-linear relationship.ResultsThe study comprised 1,965 patients (50.18% were male). The 28-day and 90-day mortality rates were 20.10% and 24.48%, respectively. The TyG index exhibited a linear relationship with the 28-day mortality (Hazards ratio (HR), 1.16; 95% confidence interval (CI), 0.99–1.36) and the 90-day mortality (HR, 1.18; 95% CI, 1.02–1.37). In the TyG Q4 group, each 1 mg/dl increase was linked to a 35% rise in the risk of 28-day mortality and a 38% increase in the risk of 90-day mortality. Subgroup analyses highlighted a more substantial association between TyG index and 90-day mortality in the diabetic group.ConclusionOur findings underscore the positive association between TyG and the 28- and 90-day mortality rates in patients with CVD. This insight may prove pivotal for identifying at-risk populations and enhancing risk prediction in the clinical management of CVD.

The transformation of risk modelling in cardiac and thoracic surgery through artificial intelligence.

The transformation of risk modelling in cardiac and thoracic surgery through artificial intelligence.

RISK MANAGEMENT IN IT-BUSINESS

Introduction. The article is devoted to topical issues of analysis, assessment and management of IT-risks in the business environment. The application of appropriate technologies enables companies to identify, assess and minimize risks in the decision-making process. Such technologies may include data analysis, risk modeling, and simulation, as well as risk management strategies. Data analysis can help businesses to identify trends and predict possible risks based on previous data. Risk modelling and simulations can help companies to estimate the possibilities that certain risks will become a reality and what are the consequences. For effective business-risk management, it is also necessary to understand what risks types exist. Risks can be associated with changes in management decisions, financial, personnel and infrastructure risks, etc. An important aspect of risk management is the provision of security-related risks, which are becoming increasingly relevant in the context of growing cyber threats and other forms of cybercrimes. The purpose of the article is to reveal the types, methods of risk research in the IT-business, their management and the justification of ways and measures to prevent and avoid them. Results. The types of IT risks are considered, the activities of successful companies in the IT sphere are studied in order to identify tools for influencing the magnitude of risk. The technologies of analysis, evaluation, management of business risks for systematization and development of measures for their prevention and avoidance are studied. The theoretical and practical significance is manifested in the fact that risk analysis in the management of business processes allows enterprises to avoid negative consequences for business and increase the efficiency of financial and economic activities. The factors influencing risks in business are substantiated, an assessment of potential security threats that may arise in the process of working with IT systems is carried out. Conclusions. The conducted analysis of the activities of successful IT companies shows that reducing the level of risk can be achieved with the help of the introduction of technologies, improvement of the company's products/services, the use of insurance poles, the implementation of financial measures, and other tools. Effective risk management at the enterprise is the key to its successful operation.

A feasibility study on utilizing machine learning technology to reduce the costs of gastric cancer screening in Taizhou, China.

To optimize gastric cancer screening score and reduce screening costs using machine learning models. This study included 228,634 patients from the Taizhou Gastric Cancer Screening Program. We used three machine learning models to optimize Li's gastric cancer screening score: Gradient Boosting Machine (GBM), Distributed Random Forest (DRF), and Deep Learning (DL). The performance of the binary classification models was evaluated using the area under the curve (AUC) and area under the precision-recall curve (AUCPR). In the binary classification model used to distinguish low-risk and moderate- to high-risk patients, the AUC in the GBM, DRF, and DL full models were 0.9994, 0.9982, and 0.9974, respectively, and the AUCPR was 0.9982, 0.9949, and 0.9918, respectively. Excluding Helicobacter pylori IgG antibody, pepsinogen I, and pepsinogen II, the AUC in the GBM, DRF, and DL models were 0.9932, 0.9879, and 0.9900, respectively, and the AUCPR was 0.9835, 0.9716, and 0.9752, respectively. Remodel after removing variables IgG, PGI, PGII, and G-17, the AUC in GBM, DRF, and DL was 0.8524, 0.8482, 0.8477, and AUCPR was 0.6068, 0.6008, and 0.5890, respectively. When constructing a tri-classification model, we discovered that none of the three machine learning models could effectively distinguish between patients at intermediate and high risk for gastric cancer (F1 scores in the GBM model for the low, medium and high risk: 0.9750, 0.9193, 0.5334, respectively; F1 scores in the DRF model for low, medium, and high risks: 0.9888, 0.9479, 0.6694, respectively; F1 scores in the DL model for low, medium, and high risks: 0.9812, 0.9216, 0.6394, respectively). We concluded that gastric cancer screening indicators could be optimized when distinguishing low-risk and moderate to high-risk populations, and detecting gastrin-17 alone can achieve a good discriminative effect, thus saving huge expenditures.

CLIM-SEG: A generalizable segmentation model for heat and flood risk mapping

With the escalating impact of climate change coupled with increased urbanization, many cities will experience extreme heat events and intense flooding. Current modeling approaches often fail to incorporate high-resolution, frequently updated data sources, such as aerial imagery from web mapping platforms, limiting their effectiveness in identifying areas at risk. To address this gap, the paper presents CLIM-SEG, a novel framework for high-resolution urban heat and flood risk assessment, addressing critical gaps in current climate risk modeling. This framework integrates semantic segmentation of aerial imagery with a weighted sum approach that integrates environmental, socioeconomic, and building data to provide comprehensive risk evaluations at the census tract level. CLIM-SEG synthesize land cover data with hazard and vulnerability factors, producing risk scores ranging from 0 to 1. This low-cost and efficient framework can enable urban planners to prioritize resources for flood mitigation and heat adaptation, addressing the limitations of current approaches and contributing to the field of urban planning and climate change adaptation. The propoosed methodology incorporates a custom-curated dataset of 545 aerial images, including 145 manually annotated segmentation maps, to fine-tune advanced semantic segmentation models. The optimized Segmenter model achieves a pixel accuracy of 97.85% and an Intersection over Union (IoU) of 0.9578 for key urban features, significantly outperforming baseline models. Boston is selected to represent an ideal representation for both heat and flood risk, as the city experiences severe urban heat islands, and is susceptible to coastal and riverine flooding, with over 11,000 structures expected to be affected by 2070 due to sea level rise and increased precipitation. Results from flood and heat risk models indicate that census tracts in South End have the highest flood risk, with a weighted score value of 0.825, while census tracts in the Fenway-Kenmore neighborhood show the highest heat risk, with a score of 0.991. Both of these results have also been verified with other heat and flood risk mapping sources for Boston. The proposed framework of CLIM-SEG not only addresses the challenges faced by Boston but also has the potential to be scaled to other urban areas dealing with the impacts of climate change, providing a valuable tool for risk assessment and decision-making in the face of a changing climate.

СОВРЕМЕННЫЕ МЕТОДЫ ПОДДЕРЖАНИЯ УСТОЙЧИВОГО РАЗВИТИЯ ОРГАНИЗАЦИИ

This article highlights the relevance of the study of modern economic factors contributing to the sustainable development of the organization. The work focuses on the analysis of key factors such as improving the efficiency of resource management through the use of energy-efficient technologies, improving the image and reputation of the organization through CSR, compliance with regulatory requirements through the introduction of an environmental management system, access to new markets through the development of innovations, reducing risks through risk management. It also considers such factors as strategic planning, attracting and retaining talent through the development of personnel and corporate culture and increasing investment attractiveness through partnerships with environmentally responsible investors. The paper emphasizes that despite the significant advantages of each of the methods, there are certain disadvantages, such as the high cost of implementation, the need for significant investments in social projects and staff training, as well as the risks of loss of investment in innovation. The paper also examines some mathematical methods and tools used to maintain the sustainable development of an organization, namely: linear programming, life cycle analysis, network analysis, machine learning methods, statistical analysis and risk modeling, decision support systems, modeling and forecasting of personnel needs.

ПРОБЛЕМИ ІНТЕРПРЕТАЦІЇ ТА ВИЗНАЧЕННЯ КОРУПЦІЙНИХ РИЗИКІВ У СУЧАСНІЙ НАУКОВІЙ ЛІТЕРАТУРІ ТА НОРМАТИВНО-ПРАВОВИХ АКТАХ УКРАЇНИ

Background . These days, the fight against corruption, its causes, consequences, and prerequisites has become a systemic and crucial part of state power policy and public administration activities due to the threatening nature of corruption processes in Ukraine and the risks associated with them, particularly during times of war. Such social challenges require intensification of the processes of development and practical implementation of an effective set of anti-corruption measures, as well as improvement of their institutional and regulatory support, which, in turn, requires further intensification of multi-disciplinary scientific research into conditions, determinants and incentives for the emergence and spread of corruption phenomena, development of adequate to current realities theoretical and methodological foundations for increasing the effectiveness of the state anti-corruption strategy, improving its regulatory and legal framework, optimizing the methodology and methods of assessing corruption risks by authorized bodies in the field of public administration. The aim of the article is to provide coverage of the results of the study of pressing problems, ideational-theoretical foundations, methodological basis, and specifics of disciplinary guidelines of scientific-theoretical study, as well as determination of the nature of corruption risks and features of their interpretation in the regulatory requirements of the current anticorruption legislation of Ukraine and methodological and methodical recommendations of other Ukrainian regulatory documents; justification of the need to introduce improvements to the ideational-theoretical and methodological foundations of the analysis and methodology for assessing the factors of corruption in the system of public administration bodies, as well as the correction of the connotations of the traditional conceptual apparatus of cognition and the conceptual explanation of the nature of corruption phenomena. Methods . In the process of studying the phenomenon and, above all, the problems of scientific theoretical modeling of corruption risks, epistemological possibilities of systemic, structural-functional, comparative-legal methods of cognition of social and political-legal phenomena, analytical grounds of discursive, interpretive, predictive and content analysis are used. Results . The importance of the research is substantiated, and the totality of those problems that relate to further improvement of the ideationaltheoretical and methodological foundations of the scientific study of corruption phenomena and processes is considered. Based on the principles of the methodology of scientific knowledge, systemic consideration of interpretations introduced by regulatory legal acts of Ukraine and publications of Ukrainian scientists on the risks of corruption and taking into account the specifics of their conceptual interpretations and classifications, the advantages and disadvantages of the points of view common in the modern Ukrainian corruption discourse on the conditions and mechanisms of corruption risk formation, respectively, their definitions, are revealed. The necessity of correcting the general theoretical content of the notion of "corruption risk" with the aim of its specification and normative unification with the terminology and requirements of the current legislation and regulatory documents of anti-corruption bodies (first of all, the National Agency on Corruption Prevention) has been substantiated. Conclusions . The imperfection of modern theoretical ideas about the essence of the phenomenon of corruption risks has been analyzed and characterized, a scientific comparison has been made, and the peculiarities of their interpretation and conceptual and terminological fixation have been revealed, firstly, in the current anti-corruption Ukrainian legislation and normative methodological documents, and secondly, in Ukrainian and foreign volumes on corruption, as a result of which a more universal and optimal for the modern theoretical model and practical procedures (assessment methods) definition of the notion of "corruption risk" has been formulated.

Study of the early warning model for network public opinion risks based on the interest preferences of internet users

Study of the early warning model for network public opinion risks based on the interest preferences of internet users

Bankruptcy forecast analysis: empirical study of the manufacturing subsector of Norte de Santander, Colombia

The study \"Bankruptcy Forecast Analysis in the Manufacturing Subsector of Norte de Santander, Colombia (2015 – 2022)\" offers an empirical evaluation of financial stability in this region, using the capital structure theory of Modigliani and Miller and the Altman Z-Score model for bankruptcy risks as a theoretical foundation. Through a quantitative approach that includes time series analysis, the internal and external factors that affect the financial competitiveness of the subsector are analyzed. The research highlights the contributions and limitations of capital structure theory and the Z-Score model in real practices, and discusses the integration of advanced technologies such as artificial intelligence to fine-tune financial risk predictions. Using purposeful sampling and documentary analysis, patterns are identified and business solvency is assessed, providing a detailed view of the sector\'s resilience in the face of economic challenges. This study provides valuable insights for those interested in the financial sustainability of the manufacturing sector, highlighting the importance of adapting risk management strategies to regional particularities. Its contribution to the specialized literature lies in offering an updated perspective on bankruptcy forecasting and financial management adapted to the specific context of the Colombian emerging economy.

Simulation Modeling of a Conceptual Model for Supply Chain Risks in Japan’s Automobile Industry

Simulation Modeling of a Conceptual Model for Supply Chain Risks in Japan’s Automobile Industry

Australian Youth Self-Harm Atlas: spatial modelling and mapping of self-harm prevalence and related risk and protective factors to inform youth suicide prevention strategies.

Suicide prevention strategies have shifted in many countries, from a national approach to one that is regionally tailored and responsive to local community needs. Previous Australian studies support this approach. However, most studies have focused on suicide deaths which may not fully capture a complete understanding of prevention needs, and few have focused on the priority population of youth. This was the first nationwide study to examine regional variability of self-harm prevalence and related factors in Australian young people. A random sample of Australian adolescents (12-17-year-olds) were recruited as part of the Young Minds Matter (YMM) survey. Participants completed self-report questions on self-harm (i.e., non-suicidal self-harm and suicide attempts) in the previous 12months. Using mixed effects regressions, an area-level model was built with YMM and Census data to produce out-of-sample small area predictions for self-harm prevalence. Spatial unit of analysis was Statistical Area Level 1 (average population 400 people), and all prevalence estimates were updated to 2019. Across Australia, there was large variability in youth self-harm prevalence estimates. Northern Territory, Western Australia, and South Australia had the highest estimated state prevalence. Psychological distress and depression were factors which best predicted self-harm at an individual level. At an area-level, the strongest predictor was a high percentage of single unemployed parents, while being in an area where ≥30% of parents were born overseas was associated with reduced odds of self-harm. This study identified characteristics of regions with lower and higher youth self-harm risk. These findings should assist governments and communities with developing and implementing regionally appropriate youth suicide prevention interventions and initiatives.

Geospatial modelling of forest fire risk and mitigation strategies: a case study

Geospatial modelling of forest fire risk and mitigation strategies: a case study

Rapid neighbourhood-scale modelling of urban heat risks for early-stage quantification of potential passive design mitigation strategy effectiveness: Copenhagen test case

This study employed the urban weather generator (UWG) to model the formation of the urban heat island (UHI) across the twelve (12) districts (bydele) of the City of Copenhagen (Københavns Kommune). In the face of projected future population growth, increased urban development, and rising global temperatures, it is necessary to quantify the ability of sustainable urban development objectives to mitigate the formation of UHIs. Assessment of baseline UHI formation in Copenhagen allowed for testing passive UHI mitigation strategies as defined in Copenhagen’s Kommuneplan 19 (KP19). The results of this study identified city districts experiencing higher baseline levels of urban heat and the effectiveness of the passive UHI mitigation strategies proposed in KP19. Additionally, results from this study suggest that UHI mitigation strategies are more effective combined.

Gut Microbiome and Hepatic Transcriptomic Determinants of HCC Development in Mice with Metabolic Dysfunction-Associated Steatohepatitis.

Hepatocellular carcinoma (HCC) related to metabolic dysfunction-associated steatotic liver disease (MASLD) is often diagnosed at a late stage, and its incidence is increasing. Predictive biomarkers are therefore needed to identify individuals at high risk of HCC. We aimed to characterize the gut microbiome and hepatic transcriptome associated with HCC development in female mice with hepatocyte-deletion of Pten (HepPten -). These mice present with large variations in HCC development, making them a powerful model for biomarker discovery. Sequencing of stool 16S and hepatic RNA was performed on a first set of mice. Among all liver histology parameters measured, the strongest association with microbiome composition changes was with the number of tumors detected at necropsy, followed by inflammation. The gut microbiome of mice with more than 2 tumors was enriched with Lachnospiraceae UCG and depleted of Palleniella intestinalis and Odoribacter. In contrast, hepatic transcriptomic changes were most strongly associated with tumor burden, followed by liver fibrosis. The 840 differentially expressed genes correlating with tumor burden were enriched in leukocyte extravasation and interleukin 10 receptor A (IL10RA) pathways. In addition, the abundance of Spp1-high epithelial cells is correlated with tumor burden. Association between tumor number and depletion of Palleniella intestinalis, and between tumor burden and circulating levels of C-X-C motif chemokine ligand 13 (CXCL13) and stem cell factor (SCF), was further validated in an independent set of mice. We identified microbiome components contributing to liver carcinogenesis by inducing inflammation, and changes in hepatic gene expression and hepatic cells distribution that contribute to tumor growth. Such information can be highly valuable for the development of new prevention strategies as well as of new biomarkers for risk modeling in HCC.