Funding Rateλ₯Ό νμ©ν΄ BTC μ λ¬Ό μμ₯μ κ°κ²© λ°©ν₯μ±μ λΆμνλ λ°μ΄ν° κΈ°λ° μ°κ΅¬ νλ‘μ νΈ
FUNDIλ μνΈνν μ λ¬Ό μμ₯μμ **Funding Rate(FR)**μ κ°κ²© κ°μ κ΄κ³λ₯Ό λΆμνμ¬
μμ₯ μ¬λ¦¬μ κ°κ²© λ°©ν₯μ±μ μ€λͺ
ν μ μλμ§ κ²μ¦νλ νλ‘μ νΈμ
λλ€.
κΈ°μ‘΄ μ°κ΅¬λ€μ΄ λ¨μ μκ΄κ΄κ³μ μ§μ€νλ€λ©΄,
λ³Έ νλ‘μ νΈλ λ€μμ μ€μ μ μΌλ‘ λ€λ£Ήλλ€:
- κ·Ήλ¨μ μΈ Funding Rate μν©μμμ κ°κ²© λ°μ
- λ€μν κ·Ήλ¨κ° μ μ λ°©μ λΉκ΅
- μμ₯ μν(regime)μ λ°λ₯Έ μ νΈ μ±λ₯ λ³ν
Can extreme funding rate conditions predict short-term BTC price direction?
- Time Horizon: 4 hours
- Approach: Event-based + regime-aware analysis
- Python
- Pandas / NumPy
- Binance Futures API
- PostgreSQL (optional)
- Matplotlib / Plotly (planned)
fundi-ml/
βββ data/
β βββ raw/ # μμ§ λ°μ΄ν° (Git μ μΈ)
β βββ processed/ # μ μ²λ¦¬ λ°μ΄ν° (Git μ μΈ)
βββ reports/ # λΆμ κ²°κ³Ό
βββ src/
β βββ data_collector.py
β βββ data_processor.py
β βββ signal_evaluator.py
βββ README.md
python src/data_collector.pypython src/data_processor.pypython src/signal_evaluator.pymethod | events | accuracy | avg_4h_return | best_regime
---------------------------------------------------------------
zscore | XXX | XX% | XX% | bull_high_vol
rolling_zscore | XXX | XX% | XX% | bear_low_vol
quantile | XXX | XX% | XX% | bull_low_vol
- Funding Rate κΈ°λ° μμ₯ μν λΆμ
- κ·Ήλ¨κ° μ μ λ°©μ λΉκ΅
- Regime-aware signal evaluation
- μ¬ν κ°λ₯ν λ°μ΄ν° νμ΄νλΌμΈ
- data/ λλ ν 리λ Gitμ ν¬ν¨λμ§ μμ΅λλ€.
- μλ μ€ν¬λ¦½νΈλ‘ λ°μ΄ν° μ¬μμ± κ°λ₯:
python src/data_collector.py
python src/data_processor.py- HMM κΈ°λ° Regime Detection
- λ¨Έμ λ¬λ λͺ¨λΈ (XGBoost)
- Cross-exchange λΆμ
- μ€μκ° μκ·Έλ μμ€ν
bibisam06 : hb.jade00@gmail.com Backend Developer interested in:
- Data Analysis
- Quantitative Research
- Financial Engineering
This project explores whether funding rate extremes can act as a leading indicator
for short-term price movements in cryptocurrency futures markets.