Quantitative Finance Research Hub

This hub serves as the primary research entry point for modern, ML-driven investment management. It bridges the gap between raw data science and the rigorous causal discovery required for institutional asset allocation.

Ⅰ. Informational NLP & Sentiment Alpha

Focuses on extracting predictive utility from unstructured textual data.

* [Sentiment Analysis for Financial Markets](SentimentAnalysisForFinance) — High-fidelity signal processing, Finance-BERT vs. GPT-4o benchmarks, and the **MSC Framework** (Minimum Sentiment Connectedness).

* [Text Analysis with Data Science](TextAnalysisWithDataScience) — Mathematical foundations of text representation (Sparse TF-IDF vs. Dense Transformers).

Ⅱ. Machine Learning & Optimization

Theoretical frameworks for robust portfolio construction and factor discovery.

* [Machine Learning for Investing: A Survey](MachineLearningForInvesting) — 2026 state-of-the-art survey covering **Causal Factor Investing** and **Scientific Causal Discovery**.

* [Hierarchical Risk Parity (HRP)](MachineLearningForInvesting#2-advanced-portfolio-optimization-hrp-2025) — Analytical proofs of noise reduction and overcoming Markowitz's instability.

* [Reinforcement Learning Fundamentals](ReinforcementLearningFundamentals) — Foundations of sequential decision-making applied to trade execution.

Ⅲ. Risk & Geopolitics

Modeling systemic shocks and supply chain vulnerabilities.

* [Geopolitical Risk and Investing](GeopoliticalRiskAndInvesting) — Categorization of shocks (Kinetic vs. Semantic) and conflict-resilient portfolio construction.

* [The 2026 Iran War Shock](MachineLearningForInvesting#6-case-study-the-2026-iran-war-shock) — Case study on the "Correlation Breakdown" and how ML maintained diversification.

* [Supply Chain Risk (GNNs)](MachineLearningForInvesting#4-supply-chain-risk-graph-neural-networks-gnns) — modeling systemic risk propagation via Graph Neural Networks.

Ⅳ. Foundational Mathematics

The underlying rigorous methods driving the ML layer.

* [Topological Data Analysis](TopologicalDataAnalysis) — Using persistent homology to extract structural features from high-dimensional manifolds.

* [TimeSeriesForecasting](TimeSeriesForecasting) — Advanced temporal modeling using XGBoost lag features and Walk-Forward Validation.

* [Numerical Methods](NumericalMethods) — Engineering continuous reality into finite-precision arithmetic.