Features

Download Historical Data

Create Features

• Feature Correlation
• Flipped Features

Create Model

• AWS Sagemaker

Check Model Performance

Randomized Significance Checks

• Random Portfolios
• Monte Carlo Check
• Whites Reality Checks

Machine Learning Model Performance Statistics

• Machine learning Statistics
• Visualize model

Backtest Performance Statistics

Features

Download Historical Data

Download historical data from Polygon

https://github.com/ari99/algorithmic_trading/tree/main/1_download_data

Create features

Uses a series of historical data and over 100 technical indicators as features.

https://github.com/ari99/algorithmic_trading/tree/main/2_data_prepare https://github.com/ari99/algorithmic_trading/tree/main/3_make_model/rnn/1_prepare_returns

Feature Correlation

Check Spearman correlation of features:

https://github.com/ari99/algorithmic_trading/blob/main/2_data_prepare/services/correlationMaker.py

Flipped Features

Also create flipped data and features to compare performance of algo with normal data: https://github.com/ari99/algorithmic_trading/blob/main/2_data_prepare/flipped_test_data.ipynb

Create Model

Create RNN model from features:

https://github.com/ari99/algorithmic_trading/tree/main/3_make_model

AWS Sagemaker

Creating the model on AWS Sagemaker:

https://github.com/ari99/algorithmic_trading/tree/main/3_make_model/rnn/3_create_models/sagemaker

Check Model Performance

Uses backtest verification techniques such as Whites Reality Check and Monte Carlo to test the statistical significance of the backtest result. Generates thousands of backtests using VectorBT.

https://github.com/ari99/algorithmic_trading/tree/main/4_test_model

See what prediction values from the model we should use to trigger a long or short order:

https://github.com/ari99/algorithmic_trading/blob/main/4_test_model/1_compare_long_short_mins.ipynb

Randomized Significance Checks

https://github.com/ari99/algorithmic_trading/blob/main/4_test_model/3_random_checks.ipynb

Random Portfolios

Compare performance with random portfolios

Monte Carlo Check

Monte Carlo check for statistical significance of backtest results

Whites Reality Check

White's Reality check for statistical significance of backtest results

Machine Learning Model Performance Statistics

Machine learning Statistics

https://github.com/ari99/algorithmic_trading/tree/main/5_model_ml_stats

Visualize model

https://github.com/ari99/algorithmic_trading/blob/main/5_model_ml_stats/visualize_model.ipynb

Backtest Performance Statistics

https://github.com/ari99/algorithmic_trading/tree/main/6_model_bt_stats

Paper trade using Alpaca

https://github.com/ari99/algorithmic_trading/tree/main/4_test_model/paper