Gretl package for generating datasets for cross-validation purposess
Learning parameters of a prediction function is a major objective. In order to avoid over-fitting, the model is trained on a different data sample (the 'training set') and evaluated out-of-sample on a so called 'test set'. This approach is known as cross-validation (CV).
This package comprises different approaches of dividing the raw data set into a training and test set, respectively, depending on the nature of the underlying data. Name the following approaches are supported:
- k-fold CV
- Leave-One-Out
- recursive moving-window (expanding window width)
- rolling moving-window (fixed window width)
An nice overview can be accessed here: https://scikit-learn.org/stable/modules/cross_validation.html
The separated data sets are stored in array of matrices, which can be read-in for the original model training and evaluation.
Written by: Artur Tarassow (atecon@posteo.de) Project page: https://github.com/atecon/cv_data_splitter