XGB-2 Classification (Taiwan Credit)

Experiment initialization and data preparation

from piml import Experiment
from piml.models import XGB2Classifier

exp = Experiment()
exp.data_loader(data="TaiwanCredit", silent=True)
exp.data_summary(feature_exclude=["LIMIT_BAL", "SEX", "EDUCATION", "MARRIAGE", "AGE"], silent=True)
exp.data_prepare(target="FlagDefault", task_type="classification", silent=True)

Train Model

exp.model_train(model=XGB2Classifier(), name='XGB2')

# Train Model with monotonicity constraints on PAY_1
exp.model_train(model=XGB2Classifier(mono_increasing_list=("PAY_1", )), name="Mono-XGB2")

Evaluate predictive performance of XGB2

exp.model_diagnose(model='XGB2', show='accuracy_table')

          ACC      AUC      F1 LogLoss   Brier

Train  0.8219   0.7978  0.4759  0.4196  0.1316
Test   0.8290   0.7728  0.4797  0.4252  0.1319
Gap    0.0071  -0.0251  0.0038  0.0057  0.0004

Evaluate predictive performance of Mono-XGB2

exp.model_diagnose(model='Mono-XGB2', show='accuracy_table')

          ACC      AUC      F1 LogLoss   Brier

Train  0.8216   0.7968  0.4780  0.4202  0.1318
Test   0.8288   0.7739  0.4816  0.4249  0.1320
Gap    0.0072  -0.0229  0.0036  0.0047  0.0001

Global effect plot for PAY_1 of XGB2

exp.model_interpret(model='XGB2', show="global_effect_plot", uni_feature="PAY_1", original_scale=True, figsize=(5, 4))

Global effect plot for PAY_1 of Mono-XGB2

exp.model_interpret(model='Mono-XGB2', show="global_effect_plot", uni_feature="PAY_1", original_scale=True, figsize=(5, 4))

Effect importance

exp.model_interpret(model='Mono-XGB2', show="global_ei", figsize=(5, 4))

Feature importance

exp.model_interpret(model='Mono-XGB2', show="global_fi", figsize=(5, 4))

Local interpretation by effect

exp.model_interpret(model='Mono-XGB2', show="local_ei", sample_id=0, original_scale=True, figsize=(5, 4))

Local interpretation by feature

exp.model_interpret(model='Mono-XGB2', show="local_fi", sample_id=0, original_scale=True, figsize=(5, 4))