yasc.scorecard.
mono_bin
(Y, X, n=20, precision=3, duplicates='raise')¶Generate monotonous bins.
A series of labels.
The series to bin, it should be of numeric type.
Number of quantiles, by default 20
The precision at which to store and display the bins labels, by default 3
Argument used by pandas.qcut()
, by default “raise”
Descriptive statistics of binning.
Examples
>>> from yasc.data import german_data
>>> from yasc.scorecard import mono_bin
>>> import pandas as pd; pd.set_option('max_columns', None)
>>> data = german_data()
>>> mono_bin(data.Creditability, data.DurationInMonth, duplicates='drop')
>>> bin_stat
min max bad_count good_count total bad_rate good_rate \
Bucket
(3.999, 12.0] 4 12 76 283 359 0.253333 0.404286
(12.0, 24.0] 13 24 122 289 411 0.406667 0.412857
(24.0, 72.0] 26 72 102 128 230 0.340000 0.182857
woe iv iv_sum bins
Bucket
(3.999, 12.0] -0.467416 0.070558 0.168117 [-inf, 12, 24, 72, inf]
(12.0, 24.0] -0.015108 0.000094 0.168117 [-inf, 12, 24, 72, inf]
(24.0, 72.0] 0.620240 0.097466 0.168117 [-inf, 12, 24, 72, inf]