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.

The significance test for Granger causality of financial risks in stock indices

🏦 Machine Learning system for credit default prediction using a RandomForestClassifier. Features an end-to-end pipeline including synthetic financial data generation, robust preprocessing (ColumnTransformer), and comprehensive evaluation with ROC-AUC and Confusion Matrices.

A comprehensive implementation of the ID3 Decision Tree algorithm from scratch for financial risk assessment, featuring custom entropy calculations, information gain optimization, and detailed data preprocessing.

Data-driven optimization of Teradyne’s excess inventory approval process using Python, lead-time adjusted demand modeling, and financial risk analysis to improve capital efficiency and reduce excess spend.

A股个人择时选股系统 — 三级择时漏斗 × 财务排雷 × 量化策略引擎 | FastAPI + Next.js 16 + SQLite

Global Systemic Fragility Observatory 2026 · Markov Model + Monte Carlo · Synthetic Bubble · 10 non-conventional systemic risk indicators · MCSH Framework · semanticfieldscience.org

Loan default risk EDA — End-to-end exploratory data analysis on imbalanced financial data (11.39:1 ratio). Univariate and bivariate segmentation by income, age, education, and organization type.

Credit-Scoring-Analysis

Built a multi-horizon (1–5 year) bankruptcy early-warning system on 43K+ firms using 64 financial ratios. Compared Altman-style linear models (ROC-AUC ~0.71) with XGBoost (ROC-AUC up to 0.97), adding temporal regime and cost-sensitive threshold analysis.

Institutional-grade early warning system for systemic deleveraging events

Monte Carlo based portfolio analytics engine with portfolio optimization, efficient frontier, and institutional risk metrics.

bankALM — Public-data U.S. bank fragility atlas. Transparent deposit-run, ALM-mismatch, and composite fragility scores for 5,000+ FDIC-insured banks.

This is a front-end system for systematic financial risk data visualization. It provides comprehensive monitoring and analysis of financial market pressure indicators across multiple dimensions including bond market, foreign exchange market, stock market, derivatives market, and money market.

A Power BI fraud detection system analyzing 6.4 million transactions. Identified $12B in financial losses, detected "Night Owl" attack patterns (2 AM–6 AM), and uncovered critical security loopholes in mobile money transfers.

Power BI dashboard analyzing loan default risk using borrower financial data and risk indicators.

Predictive modeling system for identifying high-risk financial transactions using time-based validation.

Power BI dashboard analyzing loan default risk using borrower financial data and risk indicators.

Python framework for multi-asset portfolio market risk analysis including rolling beta, VaR/CVaR, Monte Carlo simulation, stress testing, and risk decomposition.