2 alternatives explored:
- randomforest: the requirement 'Perform feature selection and remove unnecessary features' indicated a randomforest to me, and I had not seen the starter code, and as such I built this model: randomforest.ipynb
- startercode: svm.ipynb
Scores (test):
- Randomforest: 0.90
- Support Vector Machine: 0.84
- GridSearch: 0.88