C9_inprogress

Question 1

feature normalization

Answer 1

• scaling: x’ = (x - min(x)) / (max(x) - min(x))

- standardization: x’ = (x - x_gem) / sd

Question 2

dimensionality reduction

Answer 2

• select top-k most important features (use Mutual Information or Chi-square between classes and features) or use frequency as criterion
• use matrix transformation technique