University of Udine
Archivio istituzionale della ricerca - Università degli Studi di UdineNot a member yet
91472 research outputs found
Sort by
Explainable Machine Learning for the estimation of default probabilities of Italian Smes
No full textSophisticated Machine Learning (Ml) models have seen to increase predictive accuracy of linear regression models in the context of credit risk modelling. Nevertheless, linear regression models remain popular in the credit risk industry, because of the lack of transparency of Ml models. In this study, we propose a way to interpret, tuning and ex- tract default probabilities from Ml technologies in the context of credit risk. Using a sample of Italian Small and Medium sized Enterprises’ (Smes), we show how much and why Ml models increase predictions and precision of default probabilities
The L1-relaxed area of the graph of the vortex map: Optimal upper bound
No full textWe compute an upper bound for the value of the L1-relaxed area of the graph of the vortex map u: Bl(0) ⊂R2 → R2 u(x):= x/|x|, x ≠ 0, for all values of l > 0. Together with a previously proven lower bound, this upper bound turns out to be optimal. Interestingly, for the radius l in a certain range, in particular l not too large, a Plateau-type problem, having as solution a sort of catenoid constrained to contain a segment, has to be solved
Improved Stochastic SPAD Quenching Model Including Build-Up Field Effect
No full textWe present an improved stochastic model for simulating avalanche quenching in Single-Photon Avalanche Diodes (SPADs), incorporating the build-up field effect caused by space-charge accumulation during avalanche events. Built upon a one-dimensional Monte Carlo framework, our model introduces a correction term that accounts for the transient electric field induced by the charge dipole, derived from self-consistent Advection-Diffusion Monte Carlo (ADMC) simulations. This hybrid approach preserves the computational efficiency of fast stochastic methods while capturing key physical effects often neglected in simplified models. The inclusion of the build-up field is shown to impact significantly the simulated voltage and current transients, enabling more accurate prediction of quenching behavior, especially in cases of delayed or failed quenching. Comparison with full ADMC simulations shows strong agreement, validating the accuracy and relevance of the correction scheme. The proposed model provides a practical and scalable solution for SPAD simulation and design, particularly in applications requiring precise timing and high-frequency operation
Consistent and scalable composite likelihood estimation of probit models with crossed random effects
No full textEstimation of crossed random effects models commonly incurs computational costs that grow faster than linearly in the sample size, often as fast as, making them unsuitable for large datasets. For non-Gaussian responses, integrating out the random effects to obtain a marginal likelihood poses significant challenges, especially for high-dimensional integrals for which the Laplace approximation may not be accurate. In this article we develop a composite likelihood approach to probit models that replaces the crossed random effects model with some hierarchical models that require only one-dimensional integrals. We show how to consistently estimate the crossed effects model parameters from the hierarchical model fits. We find that the computation scales linearly in the sample size. The method is illustrated by applying it to approximately five million observations from Stitch Fix, where the crossed effects formulation would require an integral of dimension larger than
Ability of triglyceride-glucose indices to predict metabolic dysfunction associated with steatotic liver disease in pediatric obesity
No full textObjectives: We aimed to evaluate the best cut-off of alanine aminotransferase (ALT) to predict the metabolic dysfunction-associated steatotic liver disease (MASLD) in youths with overweight/obesity (OW/OB) and analyze the performance of the triglyceride-glucose index (TyG) or its association with anthropometric variables compared to ALT. Methods: This multicenter, cross-sectional study analyzed data of 2813 youths (1463 boys and 1350 girls) aged 6–17 years recruited in 10 Italian centers for the management of pediatric OB. Exclusion criteria were: body mass index (BMI) Z-score >5, diabetes, secondary obesity, triglycerides (TG) ≥ 400 mg/dL. MASLD was defined on the presence of hepatic steatosis at abdominal ultrasound, in combination with at least one cardiometabolic risk factor. The performance of ALT and TyG-derived indices for MASLD diagnosis was assessed using the receiver operation curve and the area under the curve (AUC) with the 95% confidence intervals were obtained. Results: MASLD was observed in 1278 youths (45.1%). The global AUC of ALT was 0.661 (0.641–0.682), p < 0.0001. Using Youden's index the best cut-off of ALT was ≥26 IU/L in boys and ≥22 IU/L in girls. The AUCs of TyG-BMI, TyG-WC, and TyG-ALT were 0.678 (0.659–0.698), 0.659 (0.639–0.679), and 0.666 (0.645–0.686), respectively. Using Delong's test none TyG-derived index showed a better performance than ALT alone. Conclusions: This study demonstrates that the values of ALT ≥ 26 IU/L in boys and ≥22 IU/L in girls can be used for the screening of MASLD in youths with OW/OB, while the combination of TyG-derived indices is not superior to ALT alone
Bertiolo: persone, famiglie, professioni e mestieri
No full textIl contributo prende in esame le dinamiche demografiche e la struttura socio-economica della comunità di Bertiolo, comune del Medio Friuli. L'arco temporale interessato concerne il XX secolo e il primo quarto del XXI
A New Model for Bone Health Management in Postmenopausal Early Breast Cancer Patients Undergoing Adjuvant Endocrine Therapy—The Predict & Prevent Project
No full textObjective: The Italian Drug Agency (AIFA) Determination n. 589 of 2015 (Note 79) establishes that the use of bisphosphonates or denosumab is necessary for the primary prevention of bone fractures in postmenopausal patients with early breast cancer (EBC) undergoing adjuvant endocrine therapy (ET). Since adherence to the 2015 AIFA recommendation was still very low in 2019, a new bone health management model was identified to improve adherence to this recommendation. Methods: The aim of this project (Predict & Prevent) was to increase the percentage of patients with early breast cancer (EBC) with hormone receptor-positive (HR+) tumors treated. The project identified a new bone health model of management including the following: training of breast multidisciplinary teams and bone health specialists; presentation and implementation of this model in cancer centers; evaluation, at baseline and 12 months after the implementation of the project, of two key performance indicators (KPIs): rate of HR+ EBC patients assessed for bone health within 30 days from the start of adjuvant ET (KPI-1) and rate of HR+ breast cancer patients receiving bisphosphonates or denosumab within 90 days from the start of adjuvant ET (KPI-2). The primary endpoints of this study were the assessment of the rates of the two key performance indicators (KPIs) 12 months after the start of the project (T3) in comparison with the rates recorded at time 0 (T0) in each participating cancer center and the bone fracture rates at 5 years. In this first analysis, we reported the rates of two KPIs 12 months after the start of the new model (T3) and the comparison with the rates recorded at time 0 (T0) in each participating cancer center, to assess whether these percentages had increased after the implementation of the new organizational model. The rates of bone fractures will be evaluated after five years from implementation of this project in every cancer center. Results: From 2020 to 2022, 10 Italian cancer centers were involved in this project. As of September 2023, 9 cancer centers reported rates relative to two KPIs assessed in each hospital. In 6 hospitals (Negrar, Brescia, Bergamo, Aviano, Turin, Rome), the rates relative to KPI-1 and to KPI-2 increased progressively from time T0 (at baseline) to time T3 (after 12 months from the start of the project), due to training of multidisciplinary teams and implementation of a new bone health management model. In the other three cancer centers (Ancona, Genoa, Naples), where the rate of evaluation of bone health (KPI-1) and the indication for bisphosphonates/denosumab (KPI-2) in HR+ EBC patients were already high at time T0, the rates remained high even after 12 months from the start of this project. Conclusions: After 12 months from the implementation of this new organizational model of bone health management, an increase in the rate of postmenopausal HR+ EBC patients on adjuvant ET assessed for bone health and the rate of patients treated with bisphosphonate/denosumab were reported in six out of nine cancer centers. In the other three cancer centers, where the rates were high at baseline, the rates also remained high after 12 months from the new model implementation. This new model should be adopted in all cancer centers to allow adequate management of bone health in all postmenopausal HR+ EBC patients undergoing adjuvant ET, with the ultimate goal of reducing the rate of bone fractures in these patients in subsequent years
Diagnosed Patients With Chronic Hepatitis B and Delta Virus in Italy in 2024: An Estimation From a National Real-World Database
No full textBackground and Aims: Hepatitis B (HBV) and Hepatitis Delta virus (HDV) infection have undergone significant changes in Italy over the past few decades, but reliable and updated prevalence of chronic hepatitis B (CHB) and Delta (CHD) data are lacking. The aim of the study was to describe the epidemiology of CHB and CHD in Italy in 2024, based on real-world data. Methods: The number of patients with a healthcare expenditure exemption for CHB (016.070.32) and CHD (016.070.33) was obtained from 21 Regional Health Authorities. To understand how many CHB or CHD patients did not have these specific exemptions, a survey was conducted in 30 Gastroenterology, Hepatology and Infectious Diseases Units across the country. Results: Health Authorities data reported 67 514 and 5216 subjects with an exemption for CHB and for CHD, respectively. However, among 6775 CHB and 504 CHD patients, only 60.3% and 55.7% of them had the specific exemption, respectively. Based on these results, we estimated 111 960 (95% CI: 109 780–114 240) CHB and 9360 (95% CI: 8690–10 150) CHD patients, with a prevalence of 0.22% and 0.019% of the adult overall population. Moreover, anti-HDV prevalence was 7.7% from this cohort. Conclusion: Our study provides a plausible estimate of the current number of adult patients diagnosed with CHB and CHD in Italy and may be considered the basis for decision-making health policies