mlresearch.utils.load_datasets

mlresearch.utils.load_datasets(data_dir, prefix='', suffix='', target_exists=True, **read_csv_kwargs)

Load all datasets in a directory from sqlite databases and/or csv files.

Parameters:

data_dirstr

Data directory to be crawled.

prefixstr, default=’’

Load dataset if the file starts with the specified prefix.

suffixstr, default=’’

Load dataset if the file starts with the specified suffix.

target_existsbool, default=True

Specify wether there is a target feature. If True, it is assumed to be in the last position of the dataset.

Returns:

datasetslist

A list with nested tuples with structure (dataset_name, (X, y)).