algofairness / fairness-comparison Goto Github PK

Hi,

In which python version those this work?

using python 3.7 on Windows 10. installed with pip3 utilized

from fairness.benchmark import run
Available algorithms:
SVM
GaussianNB
LR
DecisionTree
Kamishima
Calders
ZafarBaseline
ZafarFairness
ZafarAccuracy
Kamishima-accuracy
Kamishima-DIavgall
Feldman-SVM
Feldman-GaussianNB
Feldman-LR
Feldman-DecisionTree
Feldman-SVM-DIavgall
Feldman-SVM-accuracy
Feldman-GaussianNB-DIavgall
Feldman-GaussianNB-accuracy

run()
Datasets: '['ricci', 'adult', 'german', 'propublica-recidivism', 'propublica-violent-recidivism']'

Evaluating dataset:ricci
Sensitive attribute:Race
Algorithm: SVM
supported types: {'numerical', 'numerical-binsensitive'}
C:\Users\kevin\AppData\Local\Programs\Python\Python37-32\lib\site-packages\sklearn\metrics_classification.py:1221: UndefinedMetricWarning: Recall is ill-defined and being set to 0.0 due to no true samples. Use zero_division parameter to control this behavior.
_warn_prf(average, modifier, msg_start, len(result))
C:\Users\kevin\AppData\Local\Programs\Python\Python37-32\lib\site-packages\sklearn\metrics_classification.py:1221: UndefinedMetricWarning: Recall is ill-defined and being set to 0.0 due to no true samples. Use zero_division parameter to control this behavior.

C:\Users\kevin\Documents\GitHub\fairness-comparison-master\fairness\algorithms\kamishima\KamishimaAlgorithm.py:99: UserWarning: loadtxt: Empty input file: "C:\Users\kevin\AppData\Local\Temp\tmps24_t_kk"
m = numpy.loadtxt(output_name)
run for parameters {} failed: too many indices for array
C:\Users\kevin\Documents\GitHub\fairness-comparison-master\fairness\algorithms\kamishima\KamishimaAlgorithm.py:99: UserWarning: loadtxt: Empty input file: "C:\Users\kevin\AppData\Local\Temp\tmp86i67yum"
m = numpy.loadtxt(output_name)
run for parameters {} failed: too many indices for array
C:\Users\kevin\Documents\GitHub\fairness-comparison-master\fairness\algorithms\kamishima\KamishimaAlgorithm.py:99: UserWarning: loadtxt: Empty input file: "C:\Users\kevin\AppData\Local\Temp\tmp3q3uqt1g"
m = numpy.loadtxt(output_name)
run for parameters {} failed: too many indices for array

Above error while using actual[i] in *calc_fp_fn(actual, predicted, sensitive, unprotected_vals, positive_pred) in utility.py.
Can't figure out why. The code is available on my branch.