Data complexity measures

Common api

All metrics in this repository should be implemented following functional programming paradigm. Methods signatures should follow the prototype defined in data-complexity/prototype.py. It means that the first argument should be of type pandas.DataFrame and contain features. The second argument also should be of type pandas.DataFrame but should contain target classes. Other arguments should be passed afterwards and should have default values. Returned datatype should be of type float.

Python environment

Below you can find instructions how to create environment solving necessary dependencies. If during implementation you decide to use dependency not included in requirements.txt and environment.yml add appropriate name to mentioned file.

Conda

To create conda environment execute:

conda env create -f environment.yml

To activate created environment execute:

conda activate data-complexity

Virtual environment

To create virtual environment execute:

python -m venv data-complexity
source data-complexity/bin/activate
pip install -r requirements.txt

To activate created environment execute:

source data-complexity/bin/activate