ultra::src::holdout_validation::params Struct Reference

Public Attributes
int	training_perc {70}
int	validation_perc {30}
bool	shuffle {true}
	Whether or not to shuffle the data before splitting.
bool	stratify {true}

Member Data Documentation

stratify

bool ultra::src::holdout_validation::params::stratify {true}

Some classification problems can exhibit a large imbalance in the distribution of the target classes: for instance there could be several times more negative samples than positive samples. In such cases it's recommended to use stratified sampling to ensure that relative class frequencies is approximately preserved in train and validation sets.

Remarks

stratify is ignored for symbolic regression problems.

training_perc

int ultra::src::holdout_validation::params::training_perc {70}

Percentage of the dataset used for training.

• Valid range: 1...100;
• values <= 0 default to 70;
• values > 100 are clamped to 100.

validation_perc

int ultra::src::holdout_validation::params::validation_perc {30}

Percentage of the dataset used for validation.

Valid range: 0...(100 - training_perc); if negative or too large, it is set to 100 - training_perc.

---

The documentation for this struct was generated from the following file:

• holdout_validation.h