End-to-End Python implementation of STRAPSim: a novel portfolio similarity metric from Li et al. (2025). Combines Random Forest proximity learning with residual-aware bipartite matching to quantify economic substitutability between ETF baskets. Full replication pipeline included.

An end-to-end Python implementation of Cao et al.'s (2025) HLPPL methodology for the identification of financial (asset price) bubbles. Implements 7-parameter Log-Periodic Power Law model fitting, confidence-weighted sentiment analysis, regime-dependent 'BubbleScore' fusion, and Transformer-based forecasting with a backtesting framework.

End-to-End Python implementation of Muço’s (2025) corruption measurement framework. Combines NLP pipeline (regex extraction, Porter stemming, TF-IDF), PCA-based dimensionality reduction, and fixed-effects OLS to quantify institutional quality from Brazilian audit reports. Includes supervised learning robustness checks and LOO sensitivity analysis.

End-to-End Python implementation of Devanathan et al.'s (2026) ADMM-based distributed optimization for institutional market impact mitigation. Features 3/2-power transaction cost modeling, proximal operator calculus, VAR(1) alpha generation, and 25-year walk-forward validation, via backtesting, across 434 assets. 

End-to-End Python implementation of the Heuristic-Augmented Financial Risk Index (HAFRI) framework from Abdullah et al. (2025). Processes expert surveys and financial statements to generate composite risk scores, enabling robust ranking of fiscal periods through hierarchical decomposition and consistency validation.

End-to-End Python implementation of Mo et al.'s (2025) ACT-Tensor methodology; a tensor completion framework for financial dataset imputation. Implements cluster-based CP decomposition, HOSVD factor extraction, temporal smoothing (CMA/EMA/Kalman), and downstream asset pricing evaluation. Transforms sparse data into dense machine readable data.

📊 Calculate portfolio similarity metrics to enhance ETF alignment and optimize trading strategies in quantitative finance.

End-to-End PyTorch implementation of Chang & Kim's (2026) iVDFM pipeline, Includes: amortized encoder, RegimeNet simplex embeddings, Laplace PriorNet, diagonal AR(p) companion-form dynamics, and injective MLP decoder. Trained via ELBO with reparameterized Laplace innovations. Benchmarked against iTransformer and TimeMixer baselines.

End-to-End Python implementation of Koa et al.'s (2025) novel self-explaining quantitative framework. It combines cross-modal transformers, Time-GRPO reinforcement learning, and classifier-free guidance. Trains LLMs to perform financial technical analysis using LoRA fine-tuning. Includes backtesting with Markowitz portfolio optimization.

End-to-End Python implementation of Dávila-Fernández & Sordi's (2025) methodology for FX-constrained growth modeling (in emerging markets). Features Bayesian state-space estimation via Gibbs sampling with FFBS algorithm, heterogeneous agent simulation (fundamentalists/chartists), and nonlinear dynamics analysis.

This project replicates the results of the paper "An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale". The goal is to validate the performance of Vision Transformer (ViT) on image classification tasks using CIFAR-10.

From paper to code: a rigorous Transformer implementation in TensorFlow 2 — real WMT14 data, Moses tokenizer, and causal masking done right.

End-to-End Python implementation of Wu et al.'s (2025) ICAIF'25 paper. It translates unstructured earnings press releases into quantifiable market signals. Implements oLDA topic modeling, Transformer embeddings (BERT/FinBERT/MPNET), GPT-4o interpretability, and rigorous econometric analysis.

End-to-End Python implementation of Azcue et al.'s (2025) stochastic optimal control framework for social protection policy design. Solves PDMP-based Hamilton-Jacobi-Bellman equations using analytical closed-form solutions and Monte Carlo simulation to minimize government intervention costs (through the use of cash transfers and microinsurance).

End-to-End Python implementation of the research methodology, from "Geometric Dynamics of Consumer Credit Cycles", by Sudjianto & Setiawan (2025). Implements Clifford Algebra embeddings and Linear Attention for explanatory macroeconomic analysis; i.e. economic regime analysis.

End-to-End Python replication of Iadisernia & Camassa’s LLM macroeconomic forecasting methodology (ICAIF 2025). Implements: 2,368 synthetic economist profiles, 120,000+ GPT-4o forecasts across 50 European Central Bank (ECB) SPF rounds, a rigorous ablation study with Monte Carlo & binomial hypothesis testing.

End-to-End quantitative (Python) decision support system for optimizing economic resilience against disasters. Implements updated MRIA model using multi-regional supply-use tables, three-step optimization algorithm, and comprehensive impact assessment to identify vulnerabilities from production concentration and logistical constraints.

End-to-end Python implementation of Ma et al.'s (2025) matrix-variate diffusion index models for macroeconomic forecasting. Features α-PCA factor extraction, supervised screening, and ILS estimation for high-dimensional forecasting with preserved structural information.

End-to-End Python implementation of Markov-Switching VAR framework for detecting endogenous financial fragility. Replicates Delli Gatti et al.'s (2025) methodology using EM algorithm, Hamilton filtering, and HP spectral decomposition to empirically test Minsky's Financial Instability Hypothesis in macroeconomic data.

📊 Explore regime changes and real financial cycles through Minsky's hypothesis in a nonlinear framework, enhancing macroeconomic and financial analysis.