Credit Card Default Classifier

Supervised machine learning model for classifying if clients will default based on client features and transaction history.

Preprocessing

EDA is performed and analysis is used for feature engineering. Preprocessor transformer includes pipelines with StandardScaler, OHE, FunctionTransformer.

Machine Learning Models

A baseline dummy classifer model is created based on the most frequent target value. LogisticRegression, KNN, SVM RBF, and RandomForestClassifier approaches are examined.

Hyperparameter Optimization

Hyperparameter optimization is performed on models using GridSearchCV and RandomizedSearchCV.

Feature Importance

Feature importances are identified with RFECV.

Testing the Model

The model is tested on classfication metrics and evaluated mainly on f1 score with considerations to precision and recall.