fix: Correct column name in training feature selection

- Updated the column name from "created_timestamp" to
  "created_timestamp__" in the training feature selection
  to ensure proper data handling.

Author: danchev

credit_model.py

@@ -74,7 +74,7 @@ def _get_training_features(self, loans):
         train_X = training_df[
             training_df.columns.drop(self.target)
             .drop("event_timestamp")
-            .drop("created_timestamp")
+            .drop("created_timestamp__")
             .drop("loan_id")
             .drop("zipcode")
             .drop("dob_ssn")