Overview
This package contains the code of the first article
Recorder
This package records the results in the specified folder. It is required as it performs the serialization of non-trivial data structures like numpy arrays.
App quantile regression
This package contains the code that transforms non-probabilistic forecast methods into probabilistic ones:
- Gradient Boost
- linear regression
- k-neighbors
- lightgbm
- MLP (Multilayer perceptron)
- Random forests
All data processing
This package processes and cleans the predictors and calculates the lagged values. It separate training and testing data.
Notebooks
The notebooks contained in this package feature the visualization code for the results.