Automatic optimal sequential investment decisions. Forecasts made using advanced stochastic processes with Monte Carlo simulation. Dependency is handled with vine copulas.

Code for the WSDM 2021 paper "FluxEV: A Fast and Effective Unsupervised Framework for Time-Series Anomaly Detection".

This project studies the effects of the shape parameter estimator uncertainty at different threshold levels on the value-at-risk confidence interval for quantitative risk management (QRM) using the Generalized Pareto Distribution (GPD) from the Extreme Value Theory (EVT) approach.

DPhil project: Extreme value theory and GANs to generate compound coastal hazards (wind speed + sea level pressure) from ERA5 reanalysis data over the Bay of Bengal. In development...

Pure-Python library of heavy-tailed probability distributions (Pareto, Burr, LogNormal, etc.) built from first principles.

Python package for fitting statistical models using calibrating priors.

EVT-based noise injection toolkit for evaluating time series forecasting robustness

No description provided.

Potential Height Python packages: runs the experiments for "Finding the potential height of tropical cyclone storm surges in a changing climate using Bayesian optimization"

A deep study of human longevity using demographic data (HLD, IDL) and Extreme Value Theory to assess the potential existence of a theoretical limit to human lifespan

A Rust library and command-line tool for analyzing Power-Law distributions in empirical data.

R package for estimation of elliptical extreme quantile regions

Estimate tail parameters of heavy-tailed distributions (including power law exponent gamma) in Python

GNN for spatiotemporal Forecasting using Extreme Value Theory

Find The Tail - Matlab

A specialized Python library for sparse multivariate extreme value analysis, structure learning, and robust spectral measure estimation using extremal graphical models.

This repo contains the all the files (data, coding scripts and thesis) of my Master Thesis in Mathematics, where I had the opportunity to develop a ML model from scratch.

Alpha Asymmetry in Foreign Exchange Markets: An Investigation of Exploitability — a null result paper | DAI-2605 | Dissensus AI Working Paper

Extreme Value Theory for catastrophic claim severity — GPD/GEV, profile likelihood CIs, censored MLE, ExcessGPD reinsurance pricing, Solvency II 1-in-200 (144 tests)

An empirical analysis of financial time series, exploring heavy tails, volatility clustering, and risk measure estimation for the S&P 500 and Apple Inc.

Scripts and data used for the master's thesis "Portfolio selection with ES and regular vine copulae with EVT marginals" presented at the Wirtschaftsuniversität Wien in summer 2023. Code is structured for the purposes of the thesis and may not comply with best coding standards.

Automatic EVT based thresholding for extreme value analysis using peaks over thresholds (POT)

Estimating Value at Risk (VaR) and Median Shortfall using Extreme Value Theory on daily simple returns of the NASDAQ Composite index.

An End-to-End Python implementation of Köhler et al.'s (2026) orthogonalized tail-risk framework. Combines PCA-whitening spectral decomposition with Peaks-Over-Threshold EVT to quantify extreme risks in 479-dimensional financial networks. Implements Ferro-Segers clustering, dynamic residualization, and out-of-core processing for 2.6B+ data points.

Simulations for an article about extreme quantile region estimation

Analyzes 75 years of hydrological data from New Jersey to examine flood recurrence and the impact of climate change, including exceedance probabilities and recurrence intervals for significant flood events.

Project on the subject of Risk Management in the 1st year of the Master's program in Financial Mathematics. In this project, a risk management analysis of three indices is performed.

Estimation and forecasting of volatility using Financial Timeseries with Copulas. Includes models like GARCH, EWMA and EqWMA. Market risk management using CVaR, EVT, Risk Factors and Monte Carlo Simulation.

End-to-End Python implementation of Massacci et al.'s (2025) novel Randomized Alpha Test for high-dimensional factor models. Features robust OLS estimation, Extreme Value Theory-based inference, Monte Carlo simulation engine, and rolling-window empirical analysis. Handles N>T panels with non-Gaussian, heteroskedastic returns.

Two-stage frequentist framework for fusing sparse observations with dense simulations in spatial extreme value analysis