In this notebook, a multi-store, multi-item sales forecasting model is built with a sample data using scikit-learn API's random forest regressor. To increase the success of predictions, hyper-parameter search is performed with randomized search.
Running, the below command from python or terminal (git bash, cmd, power shell, ...)
!git clone https://github.com/orkunberk/bda557_store_item_sales_forecasting
will download all snapshots for all files.