[Data analysis] Feature engineering methods (Just method)

Before start

• This post is based on kaggle machine learning course.

Missing value dealing methods

• Method 1. Drop columns with missing values

• Method 2. Fills in the missing values with some number. (Imputation) (For using Mean, Median, before, after value) (Can use SimpleImputor)

• Method 3. An extension of imputation. You can append new columns for identifying which row is done with imputation)

Categorical value

• Method 1. Drop categorical variables

• Method 2. Ordinal encoding

• Method 3. One-hot encoding. You may not use this method if a count of variable which have to encode is more than 15.