Final group project done as a part of machine learning class with the goal of classifying and predicting whether the bought vehicle will be a good or a bad buy (have some defect in the near future).
- Performed missing value imputation using various techniques
- Performed feature selection through statistical tests and embedded methods
- Created and used scikit-learn pipelines
- Transformed and prepared data to be suitable for model building
- Used different undersampling and oversampling techniques to balance the dataset
- Evaluated and compared different machine learning models using cross-validation
- Hyperparameter tuning for optimization
- Used several groups of ensemble algorithms (stacking, bagging and boosting)