A5051364218- Impact of AI and Big Data on Business and Society
- COVID-19, Geopolitics, Technology, Migration
- Regional Socio-Economic Development Trends
- Banking stability, regulation, efficiency
- Financial Distress and Bankruptcy Prediction
- Complex Systems and Time Series Analysis
- Credit Risk and Financial Regulations
- Market Dynamics and Volatility
- Bayesian Modeling and Causal Inference
- Financial Risk and Volatility Modeling
- Data Mining Algorithms and Applications
- Explainable Artificial Intelligence (XAI)
- Bayesian Methods and Mixture Models
- FinTech, Crowdfunding, Digital Finance
- COVID-19 epidemiological studies
- Stock Market Forecasting Methods
- Insurance and Financial Risk Management
- Statistical Methods and Inference
- Forecasting Techniques and Applications
- Imbalanced Data Classification Techniques
- Global Financial Crisis and Policies
- Blockchain Technology Applications and Security
- Italy: Economic History and Contemporary Issues
- Data-Driven Disease Surveillance
- Complex Network Analysis Techniques
University of Pavia
2016-2025
Xperi (Ireland)
2024
Free University of Bozen-Bolzano
2021
University of Genoa
2021
Sapienza University of Rome
2020
University of Chieti-Pescara
2020
University of Turin
2020
Bank of Italy
2020
Human Technopole
2019
Arcada University of Applied Sciences
2018
Abstract The paper proposes an explainable Artificial Intelligence model that can be used in credit risk management and, particular, measuring the risks arise when is borrowed employing peer to lending platforms. applies correlation networks Shapley values so predictions are grouped according similarity underlying explanations. empirical analysis of 15,000 small and medium companies asking for reveals both risky not borrowers a set similar financial characteristics, which employed explain...
In credit risk estimation, the most important element is obtaining a probability of default as close possible to effective risk. This effort quickly prompted new, powerful algorithms that reach far higher accuracy, but at cost losing intelligibility, such Gradient Boosting or ensemble methods. These models are usually referred “black-boxes”, implying you know inputs and output, there little way understand what going on under hood. As response that, we have seen several different Explainable...
Financial technologies, boosted by the availability of machine learning models, are expanding in all areas finance: from payments (peer to peer lending) asset management (robot advisors) (blockchain coins). Machine models typically achieve a high accuracy at expense an insufficient explainability. Moreover, according proposed regulations, high-risk AI applications based on must be "trustworthy", and comply with set mandatory requirements, such as Sustainability Fairness. To date there no...
Financial institutions are increasingly leveraging on advanced technologies, facilitated by the availability of Machine Learning methods that being integrated into several applications, such as credit scoring, anomaly detection, internal controls and regulatory compliance. Despite their high predictive accuracy, models may not provide sufficient explainability, robustness and/or fairness; therefore, they be trustworthy for involved stakeholders, business users, auditors, regulators...
Short-term forecasts of infectious disease burden can contribute to situational awareness and aid capacity planning. Based on best practice in other fields recent insights epidemiology, one maximise the predictive performance such if multiple models are combined into an ensemble. Here, we report ensembles predicting COVID-19 cases deaths across Europe between 08 March 2021 07 2022.
The growth of Artificial Intelligence applications requires to develop risk management models that can balance opportunities with risks. We contribute the development proposing a Rank Graduation Box (RGB), set integrated statistical metrics measure "Sustainability", "Accuracy", "Fairness" and "Explainability" any application. Our are consistent each other, as they all derived from common underlying methodology: Lorenz curve. validity is assessed by means their practical application both...
There is a growing concern about the sustainability of artificial intelligence, in terms Environmental, Social and Governance (ESG) factors. We contribute to debate measuring impact ESG factors on one most relevant applications AI finance: credit rating. not yet conclusive evidence whether EGS In this paper, we propose several machine learning models measure such impact, set metrics that can improve their ability do so. way, and, more generally, decisions based become sustainable.
Summary The major implementational problem for reversible jump Markov chain Monte Carlo methods is that there commonly no natural way to choose proposals since Euclidean structure in the parameter space guide our choice. We consider mechanisms guiding choice of proposal. first group based on an analysis acceptance probabilities jumps. Essentially, these involve a Taylor series expansion probability around certain canonical jumps and turn out have close connections Langevin algorithms. second...
We propose a methodology for Bayesian model determination in decomposable graphical Gaussian models. To achieve this aim we consider hyper inverse Wishart prior distribution on the concentration matrix each given graph. ensure compatibility across models, such distributions are obtained by marginalisation from conditional complete explore alternative structures hyperparameters of latter, and their consequences model. Model is carried out implementing reversible jump Markov chain Monte Carlo...
The paper proposes an explainable AI model that can be used in credit risk management and, particular, measuring the risks arise when is borrowed employing scoring platforms. applies similarity networks to Shapley values, so predictions are grouped according underlying explanatory variables. empirical analysis of 15,000 small and medium companies asking for reveals both risky not borrowers a set similar financial characteristics, which employed explain understand their score therefore,...
SPECIALTY GRAND CHALLENGE article Front. Artif. Intell., 27 November 2018Sec. Artificial Intelligence in Finance Volume 1 - 2018 | https://doi.org/10.3389/frai.2018.00001
The late-2000s financial crisis stressed the need to understand world system as a network of countries, where cross-border linkages play fundamental role in spread systemic risks. Financial models, which take into account complex interrelationships between seem be an appropriate tool this context. To improve statistical performance we propose generate them by means multivariate graphical models. We then introduce Bayesian can model uncertainty account, and dynamic provide convenient...
Peer-to-Peer lending platforms may lead to cost reduction, and an improved user experience. These improvements come at the price of inaccurate credit risk measurements, which can hamper lenders endanger stability a financial system. In article, we propose how improve accuracy peer and, specifically, those who lend small medium enterprises. To achieve this goal, augment traditional scoring methods with "alternative data" that consist centrality measures derived from similarity networks among...
Lending activities, especially for small and medium enterprises (SMEs), are increasingly based on financial technologies, facilitated by the availability of advanced machine learning (ML) methods that can accurately predict performance a company from available data sources. However, despite their high predictive accuracy, ML models may not give users sufficient interpretation results. Therefore, it be adequate informed decision-making, as stated, example, in recently proposed artificial...
Abstract A key point to assess statistical forecasts is the evaluation of their predictive accuracy. Recently, a new measure, called Rank Graduation Accuracy (RGA), based on concordance between ranks predicted values and actual series observations be forecast, was proposed for assessment quality predictions. In this paper, we demonstrate that, in classification perspective, when response binary, RGA coincides both with AUROC Wilcoxon-Mann–Whitney statistic, can employed evaluate accuracy...