Improve an ML code

Modify an ML code with accuracy above 90% (possibly overfitting) from a random forest to a better model.

Optimize the code for more realistic accuracy by selecting the best features and the most appropriate model.