Credit card default prediction
•Cleaned and analysed the demographical and financial data of 30000 account holders.
•Applied different models like logistic regression, decision tree and random forest for classification.
•Performedhyperparametertuningandanalysedtheaverage4foldcross-validatedReceiverOperating
Characteristic (ROC-AUC) score foreachmodel to achieve thebest ROC-AUC score of 0.776.
•Derivedfinancialinsightsformthemodelsothattheclientcouldincreasehisprofitthroughvariousprogrammes.