kiraplenkin / woe_scoring

Monotone Weight Of Evidence Transformer and LogisticRegression model with scikit-learn API

WOE-Scoring

Monotone Weight Of Evidence Transformer and LogisticRegression model with scikit-learn API

Quickstart

1. Install the package:

pip install woe-scoring

2. Use WOETransformer:

import pandas as pd
from woe_scoring import WOETransformer
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression

df = pd.read_csv("titanic_data.csv")
train, test = train_test_split(
    df, test_size=0.3, random_state=42, stratify=df["Survived"]
)

special_cols = [
    "PassengerId",
    "Survived",
    "Name",
    "Ticket",
    "Cabin",
]

cat_cols = [
    "Pclass",
    "Sex",
    "SibSp",
    "Parch",
    "Embarked",
]

encoder = WOETransformer(
    max_bins=8,
    min_pct_group=0.1,
    diff_woe_threshold=0.1,
    cat_features=cat_cols,
    special_cols=special_cols,
    n_jobs=-1,
    merge_type="chi2",
)

encoder.fit(train, train["Survived"])
encoder.save_to_file("train_dict.json")

encoder.load_woe_iv_dict("train_dict.json")
encoder.refit(train, train["Survived"])

enc_train = encoder.transform(train)
enc_test = encoder.transform(test)

model = LogisticRegression()
model.fit(enc_train, train["Survived"])
test_proba = model.predict_proba(enc_test)[:, 1]

3. Use CreateModel:

import pandas as pd
from woe_scoring import CreateModel
from sklearn.model_selection import train_test_split

df = pd.read_csv("titanic_data.csv")
train, test = train_test_split(
    df, test_size=0.3, random_state=42, stratify=df["Survived"]
)

special_cols = [
    "PassengerId",
    "Survived",
    "Name",
    "Ticket",
    "Cabin",
]

model = CreateModel(
    max_vars=5,
    special_cols=special_cols,
    selection_method="sfs",
    model_type="sklearn",
    gini_threshold=5.0,
    n_jobs=-1,
    random_state=42,
    class_weight="balanced",
    cv=3,
)
model.fit(train, train["Survived"])
test_proba = model.predict_proba(test[model.feature_names_])

print(model.coef_, model.intercept_)
print(model.feature_names_)

woe_scoring's Issues

Refit bug

Min 2 bin for percent option in max_bins

Wrong len of bins

Add documentation

Change cat-binning

q_list = list(sorted(set(q_list)))

new_bins = [copy.deepcopy(bad_rates[0]["bin"])]
start = 1
for i in range(len(q_list) - 1):
    for n in range(start, len(bad_rates)):
        if (bad_rate_list[n] >= q_list[i]) & (
            bad_rate_list[n] < q_list[i + 1]
        ):
            try:
                new_bins[i] += bad_rates[n]["bin"]
                start += 1
            except IndexError:
                new_bins.append([])
                new_bins[i] += bad_rates[n]["bin"]
                start += 1

Change _mono_flags

Change to numpy.diff

Save scorecard func

Req model_results information

AttributeError: 'WaldTestResults' object has no attribute 'as_text'

Class CreateModel
Method fit

save_report = True

with open("./model_wald.txt", "w") as fh:
                fh.write(temp_model.wald_test_terms().as_text())

Set threshold diff_woe in _check_diff_woe as parameter

Bug in mono check

while (_mono_flags(bad_rates) is False) and (len(bad_rates) > 2):

Add parallel and Multicore Computing

Add default path for saving reports

Copy of dataset

Bug calc woe

(good + 0.5 / all_good) / (bad + 0.5 / all_bad)

Create model should return new feature_names_

Rename "WOE_" to prefix

Bins refactoring

bins.extend(
np.nanquantile(x, quantile / max_bins, axis=0)
for quantile in range(1, max_bins)
)

->

_, nbins = pd.qcut(x, q= max_bins, retbins=True)
bins += list(nbins)

Fix check_mono func

Num binning ->
while (_mono_flags(bad_rates) is False) and (len(bad_rates) > 2): and to top

_mono_flags ->
return True in [positive_mono_diff, negative_mono_diff]

Bug with checking Wald threshold

retrain = False
to_drop = []
for i, pvalue in enumerate(self.model.wald_test_terms().table["pvalue"]):
    if pvalue > 0.05:
        to_drop.append(self.model.wald_test_terms().table.index[i])
        retrain = True
if retrain:
    self.feature_names_ = [feature for feature in self.feature_names_ if feature not in to_drop]
    self.model = sm.Logit(y, sm.add_constant(x[self.feature_names_])).fit()

Save reports

434 if self.save_reports:
435 try:
436 with open(

AttributeError: 'CreateModel' object has no attribute 'save_reports'

max_bins = -1

Add max_bins=-1

Optimize cat_bining func

Calculate all bad and all good without nan

all_bad = y[~pd.isna(x)].sum()
all_good = len(y[~pd.isna(x)]) - all_bad

Add compare data type for cat features in refit func

try: X = X.astype(float) except ValueError: X = X.astype(object)

Fix model.predict_proba(oot)

Add lxml to requirements

Add optim parameters to logreg

Refactor transform method

Position of missing bin if only two bins