README.md

1. Kaggle competition: G-Research Crypto Forecasting

2022-02-22

• 1. Kaggle competition: G-Research Crypto Forecasting
• 2. Github file info (updating)
  • 2.1 Statistical Inference
  • 2.2 Neural Network Forecasting
  • 2.3 XGBoost/LigntGBM Forecasting (Core)
  • 2.4 Trading Strategy (backtest simulation)

The simultaneous activity of thousands of traders ensures that most signals will be transitory, persistent alpha will be exceptionally difficult to find, and the danger of overfitting will be considerable. In addition, since 2018, interest in the cryptomarket has exploded, so the volatility and correlation structure in our data are likely to be highly non-stationary. The successful contestant will pay careful attention to these considerations, and in the process gain valuable insight into the art and science of financial forecasting.

2. Github file info (updating)

• ./script/: scripts folder contains the utility script for LSTM, XGBoost, Paper presentation, general statistical inference.
• ./trainedNN: store trained RNN/LSTM models(hidden).
• ./trainedXGB: store trained XGBoost models(hidden).

2.1 Statistical Inference

• statistical-analysis-additional.ipynb: explore the crypto market by:
  • frequency manipulation script/morestates.py-> ts_with_frequency()
  • autocorrelation
  • time-series decomposition script/morestates.py> ts_decomp()
  • stationarity tests Augmented Dickey-Fuller test

2.2 Neural Network Forecasting

• RNN_forecasting.ipynb: Do RNN forecasting on the single crypto BTC OHLCV.
  • Tensorflow
  • Keras: tensorflow.keras
  • RNN-LSTM: tensorflow.keras.layers.LSTM
• my-crypto-lstm.ipynb: Manage the Kaggle competition with LSTM forecasting on the 14 cryptocurrencies returns simulaneously.

2.3 XGBoost/LigntGBM Forecasting (Core)

• crypto-xgb-paramstune.ipynb: tune xgboost hyperparameters and feature parameters.
  • xgb.cv
  • sklearn.model_selection.ParameterSampler
• crypto-xgb-scoring.ipynb: Use the weighted correlation metric to score the models' prediction as the Competition.
• crypto-xgb-api.ipynb: Debug the kaggle competition submission API.

2.4 Trading Strategy (backtest simulation)

• MA_cross_strategy.ipynb: Moving Average Crossing example of trading strategy, backtesting and evaluation.
  • generate strategy signals: script/strategy.py -> mac()
  • backtest: script/backtest.py -> bt()
  • evaluation:
    • Sharpe ratio
    • Maximum Drawdown
    • Compound Annual Growth Rate (CAGR)
    • distribution of returns
    • trade-level metrics
• cointegration_strategy.ipynb: Cointegration Strategy
  • statsmodels.tsa.stattools.coint
  • $y_{t}-\beta x_{t}=u_{t}$