Quantlab-Finder-Strategy

QuantLab finder-strategy adalah pipeline+UI trading kuantitatif berstandar institusional. Membangun infrastruktur data dan validasi statistik yang robust.

Visi

Membangun sistem trading kuantitatif yang:

• Lolos seleksi QuantConnect Alpha Streams
• Menghasilkan strategi dengan PSR > 95% dan korelasi pasar < 0.3
• Berbasis Python murni.

Filosofi: "Membangun SEBAB yang kuat, AKIBAT (profit) akan datang sendiri." Dengan:

• 6 FASE development
• 15 steering files
• Multi-engine backtest → VectorBT (FAST SCREENING Vectorized) from https://github.com/polakowo/vectorbt → Nautilus (FAST VALIDATION EVEN DRIVEN) from https://github.com/nautechsystems/nautilus_trader → LEAN (ready to validate real account) from https://github.com/QuantConnect/Lean
• Target: PSR > 95%, Correlation < 0.3, Max DD < 20%

Libraries

• Data: NumPy, Pandas, PyArrow, ArcticDB
• Scientific: SciPy, Statsmodels, Arch
• ML: Scikit-learn, LightGBM, XGBoost, hmmlearn
• Technical Analysis: TA-Lib (150+ indicators, C-optimized)
• Deep Learning: PyTorch (optional)

Data Storage

Storage	Type	Speed	Use Case
ArcticDB	Primary	3.7x faster	All OHLCV data
PyArrow	Backend	-	Required by ArcticDB

Backtest Engines

Engine	Library	Purpose
VectorBT	vectorbt	Fast screening (1000+ ideas)
Nautilus	nautilus_trader	Realistic validation (Top 50)
LEAN	QuantConnect	Alpha Streams submission

Overview Diagram

                          ┌─────────────┐
                          │ Lyer controler│
                          │  (Dash UI)  │
                          └──────┬──────┘
                                 │
                ┌────────────────┼────────────────┐
                │                │                │
          ┌─────▼─────┐   ┌─────▼─────┐   ┌─────▼─────┐
          │ Research  │   │ Backtest  │   │ Deployment│
          │ Notebooks │   │  Engines  │   │  (QC/QNT) │
          └─────┬─────┘   └─────┬─────┘   └─────┬─────┘
                │               │               │
                └───────────────┼───────────────┘
                                │
          ┌─────────────────────▼─────────────────────┐
          │              CORE LAYER                    │
          ├─────────────────────────────────────────────┤
          │  ┌──────────┐ ┌──────────┐ ┌──────────┐   │
          │  │  Data    │ │ Feature  │ │  Signal  │   │
          │  │  Engine  │ │  Engine  │ │  Engine  │   │
          │  └──────────┘ └──────────┘ └──────────┘   │
          │  ┌──────────┐ ┌──────────┐ ┌──────────┐   │
          │  │Validation│ │Portfolio │ │   Risk   │   │
          │  │  Engine  │ │  Engine  │ │  Engine  │   │
          │  └──────────┘ └──────────┘ └──────────┘   │
          └─────────────────────┬─────────────────────┘
                                │
          ┌─────────────────────▼─────────────────────┐
          │              DATA LAYER (ArcticDB)         │
          ├─────────────────────────────────────────────┤
          │  ┌──────────┐ ┌──────────┐ ┌──────────┐   │
          │  │  OHLCV   │ │ Features │ │ Signals  │   │
          │  │ Library  │ │ Library  │ │ Library  │   │
          │  └──────────┘ └──────────┘ └──────────┘   │
          └───────────────────────────────────────────┘

FASE 0: FOUNDATION

Tujuan

Membangun infrastruktur data dan validasi statistik yang robust.

Komponen

Module	Fungsi	File
ArcticStore	Time-series database (3.7x faster)	data_engine/arctic_store.py
DataManager	Unified data loading	data_engine/data_manager.py
PSRCalculator	Probabilistic Sharpe Ratio	validation_engine/psr.py
DSRCalculator	Deflated Sharpe Ratio	validation_engine/dsr.py
HurstRegime	Trending vs Mean-Reverting	signal_engine/regime/hurst.py

FASE 1: ALPHA FACTORY

Tujuan

Membuat fitur dan label yang robust untuk ML.

Komponen

Module	Fungsi	File
FractionalDiff	Stationarity dengan memory	feature_engine/fractional_diff.py
TechnicalFeatures	RSI, Bollinger, Z-Score	feature_engine/technical.py
PCADenoiser	Marcenko-Pastur denoising	feature_engine/pca_denoiser.py
TripleBarrier	Path-dependent labeling	feature_engine/labeling/triple_barrier.py
MetaLabeler	Bet sizing	feature_engine/labeling/meta_labeling.py

FASE 3: PORTFOLIO CONSTRUCTION

Tujuan

Alokasi modal dan risk management.

Komponen

Module	Fungsi	File
HRPAllocator	Hierarchical Risk Parity	portfolio_engine/hrp_allocator.py
VolatilityTargeter	Vol targeting (15%)	portfolio_engine/volatility_target.py
KellySizer	Position sizing	portfolio_engine/kelly_sizing.py
DrawdownController	DD monitoring	risk_engine/drawdown_control.py
VaRCalculator	Value at Risk	risk_engine/var_calculator.py

FASE 4: DEPLOYMENT

Tujuan

Deploy ke platform institusional.

Komponen

Module	Fungsi	File
QuantiacsAdapter	Quantiacs deployment	deployment/quantiacs/adapter.py
QuantConnectAdapter	QC Alpha Streams	deployment/quantconnect/adapter.py
PerformanceTracker	Live monitoring	deployment/monitoring/performance.py
DecayDetector	Strategy decay	deployment/monitoring/decay_detector.py
AlertSystem	Alerts	deployment/monitoring/alerts.py

FASE 5: PRODUCTION

Tujuan

Research pipeline + UI dan multi-engine backtest.

Dash UI Dashboard

QuantLab Dash UI adalah Command & Control Center untuk platform trading kuantitatif.

Quick Start

-- Demo APP in https://qlab.bamsbung.id/

---

Built with ❤️ for traders who believe in data-driven decisions.

Version 0.7.4 | 04 January 2026

📞 Support

dedy@bamsbung.id