Risk Analysis Research Articles

Machine learning for credit risk analysis across the United States

Money-Lending financial institutions face risk, which necessitates adopting a robust framework to manage it effectively. While traditional methods have been applied across the financial industry, the advent of artificial intelligence offers organizations the opportunity to utilize advanced methods to manage credit risk. This paper focuses on the application of machine learning techniques for credit risk analysis. Secondary data on information related to borrowers was extracted from Kaggle database simulating Credit Bureau data. Two ensemble models in random forest and gradient boosting were adopted for this study. The findings showed that percentage of income for loan repayment, borrower’s income, and interest rates on loans are the most important features for determining defaulters. Furthermore, the evaluation results revealed that both the random forest and the gradient boosting algorithms performed well, with F1 scores of 92.9% and 93% respectively. It was recommended that financial institutions should priorities the verification and comprehensiveness of their data, as precise data is essential for developing resilient models.

Environmental fate and safety analysis of methoxyfenozide application to control litchi and longan pests.

Litchi and longan pests significantly affect crop yield and quality. Chemical prevention and control are very effective for production; therefore, it is crucial to study fate assessment and appropriate field efficacy before pesticide application on crops to appropriately assess the health and ecological risks linked with these agents. This study conducted Good Agricultural Practice (GAP) field trials and laboratory experiments to elucidate the dissipation, terminal residues, and efficacy of methoxyfenozide on litchi and longan in six locations throughout China. To detect methoxyfenozide residues on litchi and longan, a QuEChERS/UPLC-MS/MS-based method was designed. The initial methoxyfenozide levels in litchi and longan ranged from 2.21-2.86 to 0.83-0.95mgkg-1 and indicated half-lives of 5.1-5.3 and 5.3-5.7days, respectively. After 7days of foliage treatment, the concentrations of terminal methoxyfenozide residue were 0.78-2.61 and 0.02-1.01mgkg-1, which were less than the established maximum residue limit for methoxyfenozide in litchi and longan. The chronic (acceptable daily intake = 0.0055-0.0331%) dietary intake risk analysis for methoxyfenozide in longan and litchi indicated acceptable concentrations of terminal residue for the general population. Methoxyfenozide in litchi and longan was readily degraded in first-order kinetics models, the degradation rate on longan was higher than that on litchi, and their dietary risks were negligible to consumers. Two hundred forty grams per liter of methoxyfenozide suspension concentrate (SC) represents a highly efficacious insecticidal dose to control litchi and longan pests and indicates a significant application potential as it is rapidly degraded and linked with reduced post-treatment residue levels.

The Role of Risk Analysis in EU Food Governance: Balancing Scientific Food Safety Factors and External Factors that Inform Risk Management for Healthier Food Systems

Abstract Scientific consensus links dietary choices to health outcomes, highlighting the urgency to prioritise healthy food production in food systems. At the EU level, however, defining healthy food and integrating it into food governance remains a challenge, particularly regarding the inclusion of health-related characteristics in food safety assessments. While the EU primarily relies on the risk analysis principle to address food safety concerns, the process still exhibits weaknesses that hinder its direct impact on fostering a healthier food production landscape. This review demonstrates how the risk analysis process, particularly scientific opinions in risk assessment and external factors (economic and political) in risk management, impact food governance and the healthiness of food systems. We find that while external factors play a crucial role in risk management decisions by incorporating non-food safety considerations, they often stem from an imbalance of power favouring major industry and political stakeholders. This imbalance disproportionately influences decision-making, often overshadowing nutritional and health aspects. To address these challenges, we recommend directing research towards filling knowledge gaps and exploring minority scientific findings, and separating external factors from risk management’s decision-making process, to ensure that food governance prioritises public health and healthy food production.

Systemic Risk Analysis of Safety, Progress, and Investment in the Construction of a Water Transfer Project and the Importance of Common Cause Failure

Safety, progress, and investment risks are correlated during the construction period of large-scale water transfer projects. However, previous studies have only considered individual risk factors, overlooking the potential systemic risk posed to safety, progress, and investment, as well as any underlying common cause failures. Since traditional risk analysis methods are ill-suited to addressing common cause failure, this paper’s objective was to establish a comprehensive evaluation index framework and to identify the basic events of common cause failure. To do that, we developed a risk analysis method that models common cause failure based on a Bayesian network for assessing that systemic risk. The Henan Section of the Yangtze-to-Huaihe River Water Diversion Project in China was then used as a case study. The results show that a variety of common cause failure events, such as epidemic disease, design alteration, lagged approval process, heavy rain in the flood season, renewal material and failing equipment, construction accidents, and external interference, can significantly impact the safety, progress, and investment systemic risk. Design alteration poses the greatest risk, with renewal material and failing equipment exerting the strongest influence among all common cause failure events. It is also possible to elucidate the predominant causal chains; specifically, the contributing influence of each basic failure event to the systemic risk can be clarified by adjusting their respective initial state. The failure of renewal material and failing equipment was found to significantly increase the safety risk. This study effectively simulated the complex causal relationships and uncertainties of pertinent risk factors, thereby enhancing our understanding of the systemic risk associated with safety, progress, and investment in large-scale water transfer projects.

Risk Analysis and Simulation of Large Bridge Construction Based on System Dynamics

To reduce risk incidents in large bridge construction, it is necessary to study the interaction mechanisms and dynamic changes among various risk factors during the construction of large bridges. First, the evolution mechanism of construction risks for large bridges is analyzed, and a risk factor structure system is established. Then, based on system dynamics theory, a causal loop diagram and flow–stock diagram are constructed, and system dynamics equations for each variable in the flow diagram are established to build a system dynamics model. Finally, taking a large bridge construction project as an example, the variable equations are assigned actual values, and the construction risk level is simulated and analyzed. The results show that, ➀ continuous investment in safety funds can effectively reduce the overall risk level of the system, ➁ changes in the management risk subsystem have a significant impact on the overall risk level of large bridge construction, and ➂ increasing the safety investment ratio in both the personnel risk subsystem and the management risk subsystem can effectively reduce the total risk level of the system.

Seismic analysis of gravity dam-foundation systems using stochastic spectral finite element method

Seismic analysis of dam-foundation systems is inherently complex due to uncertainties in material properties. This study investigates the effects of material randomness and heterogeneity on the seismic response of gravity dam-foundation systems using the stochastic spectral finite element method (SSFEM) based on the Karhunen Loève expansion and the polynomial chaos expansion. The Pine Plane dam is taken as an analysis example. First, the appropriate Karhunen Loève expansion modes and polynomial chaos order are determined for SSFEM by comparing them with the Monte-Carlo simulations. Subsequently, the seismic response of the dam is analyzed in the stochastic and deterministic models, respectively. Results indicate that material randomness and heterogeneity may reduce the seismic response of gravity dams, and the deterministic model may overestimate dynamic responses. Subsequently, the parametric analysis of the correlation length, which measures the spatial correlation between variables in different positions and indicates the size of heterogeneity material patches, is conducted. It is demonstrated that the correlation length would not significantly affect the mean responses, but a larger correlation length leads to greater dispersions in responses. Finally, the effects of randomness in the foundation are investigated. The heterogeneities in the foundation affect the response of the dam in the high-frequency range. Therefore, the uncertainty in material properties should be fully considered in a structural assessment and risk analysis.

Bagging Vs. Boosting in Ensemble Machine Learning? An Integrated Application to Fraud Risk Analysis in the Insurance Sector

ABSTRACT Addressing the pressing challenge of insurance fraud, which significantly impacts financial losses and trust within the insurance industry, this study introduces an innovative automated detection system utilizing ensemble machine learning (EML) algorithms. The approach encompasses four strategic phases: 1) Tackling data imbalance through diverse re-sampling methods (Over-sampling, Under-sampling, and Hybrid); 2) Optimizing feature selection (Filtering, Wrapping, and Embedding) to enhance model accuracy; 3) employing binary classification techniques (Bagging and Boosting) for effective fraud identification; and 4) applying explanatory model analysis (Shapley Additive Explanations, Break-down plot, and variable-importance Measure) to evaluate the influence of individual features on model performance. Our comprehensive analysis reveals that while not every re-sampling technique improves model performance, all feature selection methods markedly bolster predictive accuracy. Notably, the combination of the Gradient Boosting Machine (GBM) algorithm with NCR re-sampling and GBMVI feature selection emerges as the most effective configuration, offering superior fraud detection capabilities. This study not only advances the theoretical framework for combating insurance fraud through AI but also provides a practical blueprint for insurance companies aiming to incorporate advanced AI strategies into their fraud detection arsenals, thereby mitigating financial risks and fostering trust systems.

Impact of Meteorological Conditions on PM2.5 Pollution in Changchun and Associated Health Risks Analysis

The escalating concern regarding increasing air pollution and its impact on the health risks associated with PM2.5 in developing countries necessitates attention. Thus, this study utilizes the WRF-CMAQ model to simulate the effects of meteorological conditions on PM2.5 levels in Changchun, a typical city in China, during January 2017 and January 2020. Additionally, it introduces a novel health risk-based air quality index (NHAQI) to assess the influence of meteorological parameters and associated health risks. The findings indicate that in January 2020, the 2-m temperature (T2), 10-m wind speed (WS10), and planetary boundary layer height (PBLH) were lower compared to those in 2017, while air pressure exhibited a slight increase. These meteorological parameters, characterized by reduced wind speed, heightened air pressure, and lower boundary layer height—factors unfavorable for pollutant dispersion—collectively contribute to the accumulation of PM2.5 in the atmosphere. Moreover, the NHAQI proves to be more effective in evaluating health risks compared to the air quality index (AQI). The annual average decrease in NHAQI across six municipal districts from 2017 to 2020 amounts to 18.05%. Notably, the highest health risks are observed during the winter among the four seasons, particularly in densely populated areas. The pollutants contributing the most to the total excess risk (ERtotal) are PM2.5 (45.46%), PM10 (33.30%), and O3 (13.57%) in 2017, and PM2.5 (67.41%), PM10 (22.32%), and O3 (8.41%) in 2020. These results underscore the ongoing necessity for PM2.5 emission control measures while emphasizing the importance of considering meteorological parameters in the development of PM2.5 reduction strategies.

The Challenges of Blood Sample Delivery via Drones in Urban Environment: A Feasibility Study through Specific Operation Risk Assessment Methodology

In recent years, Unmanned Aircraft System (UAS) usage in the medical sector as an alternative to traditional means of goods transport has grown significantly. Even though the reduced response time achieved with UASs can be lifesaving in critical situations, their usage must comply with technological constraints such as range, speed and capacity, while minimizing potential risks. In this paper, the feasibility of a drone operation dedicated to the transport of blood samples in an urban area is studied through a safety risk analysis. The assessment utilizes the Specific Operation Risk Assessment (SORA) framework, in line with current European regulations, and extends it to define flight trajectories with minimal risk. A case study in the Helsinki urban area is used as a reference, with an exemplary case of commercial drone transportation of blood samples between the Töölö and Malmi Hospitals. By leveraging the drone performance capabilities and minimizing the risk for people on the ground, this approach demonstrates that medical delivery using drones in densely populated urban environments remains challenging. Nonetheless, it argues that the proposed method can enhance risk awareness and support the planning of feasible operations.

A Classification of Cloud Services and Their Security Assessment in the Healthcare Industry

The role of cloud-based services in healthcare settings is crucial in the 21st century. This paper discusses their advantages and tools in hospitals, clinics, and diagnostic centers, categorizing existing applications and services into data storage, computing power, networking, PaaS, SaaS, data analytics, business intelligence, and project management. It also provides an in-depth analysis of security and risk assessment issues in cloud-based services, supported by case studies. Additionally, it includes a comparative risk analysis diagram of secured versus non-secured cloud systems in healthcare. Finally, the paper concludes with recommendations for future work. Key Words: Healthcare facilities, diagnostic centers, cloud services, information technology systems, security, risk analysis.

Risk analysis of noise-induced hearing loss of workers in the automobile manufacturing industries based on back-propagation neural network model: a cross-sectional study in Han Chinese population.

This study aims to predict the risk of noise-induced hearing loss (NIHL) through a back-propagation neural network (BPNN) model. It provides an early, simple and accurate prediction method for NIHL. Population based, a cross sectional study. Han, China. This study selected 3266 Han male workers from three automobile manufacturing industries. Information including personal life habits, occupational health test information and occupational exposure history were collected and predictive factors of NIHL were screened from these workers. BPNN and logistic regression models were constructed using these predictors. The input variables of BPNN model were 20, 16 and 21 important factors screened by univariate, stepwise and lasso-logistic regression. When the BPNN model was applied to the test set, it was found to have a sensitivity (TPR) of 83.33%, a specificity (TNR) of 85.92%, an accuracy (ACC) of 85.51%, a positive predictive value (PPV) of 52.85%, a negative predictive value of 96.46% and area under the receiver operating curve (AUC) is: 0.926 (95% CI: 0.891 to 0.961), which demonstrated the better overall properties than univariate-logistic regression modelling (AUC: 0.715) (95% CI: 0.652 to 0.777). The BPNN model has better predictive performance against NIHL than the stepwise-logistic and lasso-logistic regression model in terms of TPR, TNR, ACC, PPV and NPV (p<0.05); the area under the receiver operating characteristics curve of NIHL is also higher than that of the stepwise and lasso-logistic regression model (p<0.05). It was a relatively important factor in NIHL to find cumulative noise exposure, auditory system symptoms, age, listening to music or watching video with headphones, exposure to high temperature and noise exposure time in the trained BPNN model. The BPNN model was a valuable tool in dealing with the occupational risk prediction problem of NIHL. It can be used to predict the risk of an individual NIHL.

Risk evolution from causes to consequences of engine room fires on ships by mapping bow-tie into fuzzy Bayesian network

An integrated quantitative risk analysis methodology based on the BT (bow-tie) and FBN (fuzzy Bayesian network) is developed in the present study to evaluate the risk evolution from causes to consequences of engine room fires on ships. Given limited historical statistics on engine room fire accidents, a probabilistic estimation model based on fuzzy sets is employed to determine prior probabilities of BN nodes. A case study of the engine room fire is carried out to demonstrate the soundness and applicability of the proposed methodology. Moreover, scenario analysis is implemented to assess how safety barriers can mitigate escalation of consequences triggered by engine room fires, while sensitivity analysis ranks critical hazards and quantifies correlations between different consequences and risk events within engine room fires. The results indicate that mechanical faults, improper bunker operation, improper maintenance, circuit overload or short circuit and insulation aging or damage are the main direct risk events triggering engine room fires on ships. The effectiveness of safety barrier (fire extinguishing devices) can play a decisive role in preventing further escalation of controlled fire incidents.

A Novel Security Risk Analysis Using the AHP Method in Smart Railway Systems

Transportation has an essential place in societies and importance to people in terms of its social and economic aspects. Innovative rail systems need to be integrated with developing technologies for transportation. Systemic failures, personnel errors, sabotage, and cyber-attacks in the techniques used will cause a damaged corporate reputation and revenue losses. In this study, cybersecurity attack methods in smart rail systems were determined, and cyber events occurring worldwide through these technologies were analyzed. Risk analysis in terms of transportation safety in smart rail systems was determined by considering the opinions of 10 different experts along with the Analytic Hierarchical Process (AHP) performance criteria. Informatics experts were selected from a group of people with at least 5–15 years of experience. According to these risk analysis calculations, cybersecurity stood out as the most critical security risk at 27.74%. Other risky areas included physical security, calculated at 14.59%, operator errors at 16.20%, and environmental security at 10.93%.

Evaluation of phosphate insecticides and common herbicides: monitoring and risk assessment in water treatment plant, distribution networks, and underground water wells

ABSTRACT Despite the negative effects that the use of pesticides (such as herbicides and insecticides) has on human health and water resources, a significant portion of the world's agricultural production depends on them. The purpose of this study was to determine selected residual concentrations of pesticides (diazinon, ethion, malathion, alachlor, methyl-parathion, trifluralin, atrazine, chlorpyrifos, and azinphos-methyl) in samples from Shiraz potable water sources. For this purpose, water treatment plant, groundwater wells, treated surface water, and a mixture of groundwater and treated surface water were taken. In addition, statistical and risk analyses (carcinogenic and non-carcinogenic) were used. According to the results, chlorpyrifos with 84.4% had the highest removal efficiency and methyl-parathion with 10% had the lowest removal rate in the Shiraz water treatment plant process. The highest mean concentration was related to azinphos-methyl (1.5 μg/L) and chlorpyrifos (0.59 μg/L) in the groundwater samples. All measured compounds in water source samples were below standard levels, except for chlorpyrifos and azinphos-methyl, which were reported in groundwater above the limit recommended by the Environmental Protection Agency (EPA). The results showed that while the selected pesticides measured had a low non-carcinogenic risk for both adults and children, malathion and trifluralin posed a high carcinogenic risk for adults.

Investigating the Black Birth Experience: A Race-Stratified Analysis of Preterm Birth Risk and Exposure to Metropolitan Statistical Area-Level Police-Related Deaths, US 2018-2019.

Police-related violence may be a source of chronic stress underlying entrenched racial inequities in reproductive health in the USA. Using publicly available data on police-related fatalities, we estimated total and victim race-specific rates of police-related fatalities (deaths per 100,000 population) in 2018-2019 for Metropolitan Statistical Areas (MSA) and counties within MSAs in the USA. Rates were linked to data on live births by maternal MSA and county of residence. We fit adjusted log-Poisson models with generalized estimating equations and cluster-robust standard errors to estimate the relative risk of preterm birth associated with the middle and highest tertiles of police-related fatalities compared to the lowest tertile. We included a test for heterogeneity by maternal race/ethnicity and additionally fit race/ethnicity-stratified models for associations with victim race/ethnicity-specific police-related fatality rates. Fully adjusted models indicated significant adverse associations between police-related fatality rates and relative risk of preterm birth for the total population, non-Hispanic Black, and non-Hispanic White groups separately. Results confirm the role of fatal police violence as a social determinant of population health outcomes and inequities, including preterm birth.

Evaluating the Correlation Between Anteroposterior Diameter, Body Surface Area, and Height for Liver Transplant Donors and Recipients.

Small stature and female sex correlate to decreased deceased donor liver transplant (DDLT) access and higher waitlist mortality. However, efforts are being made to improve access and equity of allocation under the new continuous distribution (CD) system. Liver anteroposterior diameter (APD) is a method used by many centers to determine size compatibility for DDLT but is not recorded systematically, so it cannot be used for allocation algorithms. We therefore seek to correlate body surface area (BSA) and height to APD in donors and recipients and compare waitlist outcomes by these factors to support their use in the CD system. APD was measured from single-center DDLT recipients and donors with cross-sectional imaging. Linear, Pearson, and PhiK correlation coefficient were used to correlate BSA and height to APD. Competing risk analysis of waitlist outcomes was performed using United Network for Organ Sharing data. For 143 pairs, donor BSA correlated better with APD than height (PhiK = 0.63 versus 0.20). For recipient all comers, neither BSA nor height were good correlates of APD, except in recipients without ascites, where BSA correlated well (PhiK = 0.63) but height did not. However, among female recipients, BSA, but not height, strongly correlated to APD regardless of ascites status (PhiK = 0.80 without, PhiK = 0.70 with). Among male recipients, BSA correlated to APD only in those without ascites (PhiK = 0.74). In multivariable models, both BSA and height were predictive of waitlist outcomes, with higher values being associated with increased access, decreased delisting for death/clinical deterioration, and decreased living donor transplant (model concordance 0.748 and 0.747, respectively). Taken together, BSA is a good surrogate for APD and can therefore be used in allocation decision making in the upcoming CD era to offset size and gender-based disparities among certain candidate populations.

Acute Kidney Injury in the Context of COVID-19: An Analysis in Hospitalized Mexican Patients

During the COVID-19 pandemic, a considerable proportion of patients developed a severe condition that included respiratory failure, shock, or multiple organ dysfunction. Acute Kidney Injury (AKI) has been recognized as a possible cause of severe COVID-19 development. Given this, this study investigates the occurrence and consequences of AKI in Mexican patients to contribute to better knowledge and management of this problem. Methods: Using a retrospective observational cohort methodology, we investigated 313 cases from a cohort of 1019 patients diagnosed with COVID-19 at the IMSS Zacatecas General Hospital of Zone No. 1 in 2020. The prevalence of AKI was determined using the AKIN criteria based on serum creatinine levels and a detailed review of demographic characteristics, medical history, comorbidities, and clinical development. Results: The data showed a 25.30% prevalence of AKI among patients infected with severe COVID-19. Remarkably, these patients with AKI exhibited an advanced age (&gt;65 years), arterial hypertension, a higher number of white blood cells during admission and the hospital stay, and elevated levels of C-reactive protein, serum creatinine, and blood urea nitrogen (BUN). Clinically, patients with AKI had signs of prostration, pneumonia, and the requirement for ventilatory assistance when compared to those without AKI. Finally, those diagnosed with AKI and COVID-19 had a 74% death rate. Relative risk analyses indicated that age (&gt;65 years), arterial hypertension, high creatinine levels, endotracheal intubation, and pneumonia are associated with the development of AKI. On the other hand, among the protective factors against AKI, high hemoglobin levels and the consumption of statins during COVID-19 were found. Conclusions: The findings of this study underscore the significance of promptly identifying and effectively managing AKI to potentially alleviate the negative consequences of this complication within the Mexican population during COVID-19.

Risk‐based student performance prediction model for engineering courses

AbstractHigh academic failure and dropout rates in engineering courses are significant worldwide concerns attributed to various factors, with academic performance being a critical variable. This article provides a methodology to estimate the performance risk of students in engineering schools. Risk analysis is a strategy to evaluate academic success, which provides a set of methods to analyze, understand, and predict student outcomes before enrolling in specific majors or challenging college courses. This article develops a methodology to estimate fragility curves for students entering an engineering course. The fragility function concept, borrowed from the earthquake engineering field, estimates the likelihood of success in a course, given relevant student metadata, such as the grade point average, thus comprehensively addressing student performance variability. A student academic success prediction model enables instructional designers to make informed decisions. For example, fragility curves can help achieve two goals: (i) assessing the population at risk for a course to take actions to improve student success rates and (ii) assessing a course's relative difficulty based on its fragility function parameters. We demonstrate this methodology through a case study comparing the relative difficulty of two engineering courses, Statics and Solid Mechanics, at a university in Colombia. Given that Statics serves as a prerequisite for Solid Mechanics, deficiencies in the former can significantly impact student performance in the latter. The case study results reveal that Solid Mechanics poses a higher risk of academic failure than Statics, underscoring the importance of a strong foundation in prerequisite courses.

What was first and what is next in selecting device-aided therapy in Parkinson's disease? Balancing evidence and experience.

Parkinson's disease (PD) progresses with motor fluctuations emerging several years after treatment initiation. Initially managed with oral medications, these fluctuations may later necessitate device-aided therapy (DATs). Globally, various DATs options are available, including continuous subcutaneous apomorphine infusion, deep brain stimulation, levodopa-carbidopa intestinal gel, levodopa-entacapone-carbidopa intestinal gel, and subcutaneous foslevodopa/foscarbidopa infusion, each with its complexities. Hence, matching complex patients with suitable therapy is critical. This review offers practical insights for physicians managing complex PD cases. Balancing evidence and experience is vital to select the most suitable DATs, considering factors like disease stage and patient preferences. Comparative analysis of DATs benefits and risks provides essential insights for clinicians and patients. Treatment sequences vary based on availability, patient needs, and disease progression. Less invasive options like apomorphine are often preferred initially, followed by other DATs if needed. Patient selection requires comprehensive evaluations, including motor function and cognitive status. Follow-up care involves symptom monitoring and adjusting medications. Customized treatment plans are essential for optimizing PD management with DATs.

The association between myasthenia gravis and risk of fracture: a systematic review and meta-analysis.

Patients with myasthenia gravis (MG), because of their muscle weakness and exposure to corticosteroids treatment, are generally considered to be at increased risk for osteoporosis or fracture. However, clinical evidence of this issue is lacking. In this review, we systematically searched databases, including Cochrane Library, PubMed, Embase, and Airiti library from inception to the end of November 2023 for cohort studies that compared participants with MG and participants without MG for incidence of osteoporosis or fracture. We used the Newcastle-Ottawa Scale for quality assessment. In total, we included 3 studies with 34,865 participants. The pooled meta-analysis using the random effect model demonstrated no significant difference in risk of fracture in the MG group (odds ratio = 1.52; 95% confidence interval = 0.74 to 3.12; I2 = 93%; between-study variance [τ2] = 0.32) compared with that for the non-MG group. Due to limited studies, we could not perform a quantitative analysis for risk of osteoporosis. In conclusion, we found no robust evidence to support the proposition that patients with MG are at higher risk for fracture than general comparators. The explanations and underlying mechanisms of this finding remain unclear, we therefore conclude that additional studies are warranted.