agenthinkmesh/islamic_finance_ai

Islamic Finance AI — 10 Model Suite

AgenThink | Kuwait City | github.com/agentthink

World's first open-source Islamic Finance AI model library. Built with synthetic GCC data and AutoML (zero manual tuning).

Models

#	Model	Type	Score
1	Sukuk Yield Prediction	Regression	R²: 0.960
2	Shariah Stock Screener	Classification	Accuracy: 1.000
3	Zakat Optimization	Regression	R²: 0.986
4	Murabaha Credit Risk	Classification	Accuracy: 1.000
5	Halal Portfolio Optimizer	Regression	R²: 0.797
6	Takaful Claims Prediction	Regression	R²: 0.845
7	Islamic Bank Liquidity Risk	Classification	Accuracy: 0.933
8	Waqf Asset Valuation	Regression	R²: 0.995
9	GCC Fintech Fraud Detection	Classification	Accuracy: 0.974
10	SWF Asset Allocation	Regression	R²: 0.542

Usage

import joblib
model = joblib.load("model_sukuk_yield.pkl")
prediction = model.predict(X_new)

Models Included

• Sukuk Yield Prediction — Predict profit rates for Islamic bonds
• Shariah Stock Screener — AAOIFI-compliant stock classification
• Zakat Optimizer — Multi-madhab Zakat calculation engine
• Murabaha Credit Risk — Islamic mortgage default prediction
• Halal Portfolio Optimizer — Sharpe ratio optimization
• Takaful Claims Prediction — Islamic insurance claim modeling
• Islamic Bank Liquidity Risk — Basel III adapted for Islamic banks
• Waqf Asset Valuation — Islamic endowment property valuation
• GCC Fintech Fraud Detection — Region-specific fraud signals
• Sovereign Wealth Fund Allocation — Oil-adjusted asset allocation

License

MIT — Free to use, modify, and deploy

Citation

AgenThink (2025). Islamic Finance AI Suite. github.com/agentthink