Model the next day's close price using Deep Learning Yearning and Tensorflow Probability
Next steps are to:
-
Model the next day's high and low prices
-
Make a new model to use open price as a predictor
-
Expand horizon to multiple days outs of weeks out either by explicit modeling or chaining
yearning_stock_ts.R: Model training
score_yearning_model_ts.R: Model evaluation
daily_scoring.R: Pull all tickers from Tiingo, and get a predicted distribution for each ticker on a given day
daily_order_entry.R: Use the output from daily_scoring.R to make orders with Alpaca
mid_day_rebalance.R: Deploy daily mean reversion strategy using the output from daily_scoring.R and submit orders with Alpaca
sell_off.R: Create a limit order for all currently held stocks with Alpaca
stop_loss.R: Submit stop loss orders for positions based on daily scoring.
stonk_weights*: Weights for Deep Learning model.
*transform_recipe.RDS: Supplementary data transformation using the recipes package.