jnjcc / upeval

Evaluation tools for Uplift Modelling, with continuous outcome support

upeval

Evaluation tools for Uplift Modelling, with continuous outcome support. Serves as a supplement to pylift.

Usage

import pandas as pd
from upeval import *
df = pd.DataFrame({
	"Treatment": [1, 1, 0, 0, 1, 1, 0, 0, 1, 1],
	"Outcome":   [6, 7, 1, 1, 7, 8, 2, 2, 9, 10],
	"Uscore":    [0.8, 0.7, 0.8, 0.7, 0.6, 0.5, 0.6, 0.5, 0.6, 0.5]
})
upe = UpContEval(df)
upq = UpContEvalQ(df, nquantiles = 3)
print(upq.auuc_score())
upq.plot_cumsums("cumsums.png")
upq.plot_quantiles("quantiles.png", density = True)

COPYING

Copyright (c) 2021 jnjcc, Yste.org

This program is licensed under the terms of the MIT license. See COPYING for more details.