A 5-day project that aims to identify credit card defaults from a client base in Taiwan. The data has 30,000 observations along with 23 financial and demogrphic features. The dataset is skewed with only 22% are defaults.
- EDA-Credit Default: (Exploratory Data Analysis) Quick, cheap-to-complex data visualizations to examine variable relationship and create potential features
- Credit Default Project - Model building: Model tuning, training, and evaluation over 5 learning classifiers
- Results and Discussion: Summary of methodology, model performance, and discussion on limitation.