scientistmetrics is a Python package for metrics and scoring : quantifying the quality of predictions
scientistmetrics provides the option for computing one of six measures of association between two nominal variables from the data given in a 2d contingency table:
- Chi - squard test : https://en.wikipedia.org/wiki/Chi-squared_test
- Cramer's V : https://en.wikipedia.org/wiki/Cram%C3%A9r%27s_V
- Tschuprow's T : https://en.wikipedia.org/wiki/Tschuprow%27s_T
- G-test : https://en.wikipedia.org/wiki/G-test
- Phi coefficient : https://en.wikipedia.org/wiki/Phi_coefficient
- Pearson contingence coefficient : https://www.statisticshowto.com/contingency-coefficient/
scientistmetrics provides metrics for classification problem :
- accuracy score
- f1 score
- precision
- recall
- etc...
scientistmetrics provides metrics for regression problem :
- Rsquared
- Adjusted Rsquared
- Mean squared error
- etc...
scientistmetrics provides a function that gives a set of all subsets model.
Notebook is availabled.
scientistmetrics requires :
python >=3.10
numpy >=1.26.4
pandas >=2.2.2
scikit-learn >=1.2.2
plotnine >=0.10.1
statsmodels >=0.14.0
scipy >=1.10.1
You can install scientistmetrics using pip
:
pip install scientistmetrics
Duvérier DJIFACK ZEBAZE duverierdjifack@gmail.com