Modelowanie ryzyka kredytowego w R
Lore Dirick
Manager of Data Science Curriculum at Flatiron School



1) Zacząć od modelu zawierającego wszystkie zmienne (w tym przypadku 7) i obliczyć AUC
log_model_full <- glm(loan_status ~ loan_amnt + grade + home_ownership +
annual_inc + age + emp_cat + ir_cat,
family = "binomial", data = training_set)
predictions_model_full <- predict(log_model_full,
newdata = test_set, type ="response")
AUC_model_full <- auc(test_set$loan_status, predictions_model_full)
Area under the curve: 0.6512
2) Zbudować 7 nowych modeli, za każdym razem usuwając jedną zmienną, i obliczyć prognozy PD na zbiorze testowym
log_1_remove_amnt <- glm(loan_status ~ grade + home_ownership + annual_inc + age + emp_cat + ir_cat,
family = "binomial",
data = training_set)
log_1_remove_grade <- glm(loan_status ~ loan_amnt + home_ownership + annual_inc + age + emp_cat + ir_cat,
family = "binomial",
data = training_set)
log_1_remove_home <- glm(loan_status ~ loan_amnt + grade + annual_inc + age + emp_cat + ir_cat,
family = "binomial",
data = training_set)
pred_1_remove_amnt <- predict(log_1_remove_amnt, newdata = test_set, type = "response")
pred_1_remove_grade <- predict(log_1_remove_grade, newdata = test_set, type = "response")
pred_1_remove_home <- predict(log_1_remove_home, newdata = test_set, type = "response")
...
3) Zachować model z najlepszym AUC (AUC pełnego modelu: 0.6512)
auc(test_set$loan_status, pred_1_remove_amnt)
Area under the curve: 0.6537
auc(test_set$loan_status, pred_1_remove_grade)
Area under the curve: 0.6438
auc(test_set$loan_status, pred_1_remove_home)
Area under the curve: 0.6537
4) Powtarzać, aż AUC zacznie (znacząco) maleć
Modelowanie ryzyka kredytowego w R