Data Pre-Processing¶

• Identification of Dependent and Independent features
• Nominal features
• Missing values features
• Feature scaling

Identification of Dependent and Independent features¶

Understand the problem at hand.

Nominal features¶

• Features that do not have any inherent numerical value or ordinal relationship.
• Find out the number of possible values a few basic features take and make a note of it.
• Encode them with [0, 1, 2, ...] or something similar.

Missing values features¶

We can do either of the three given methods to get rid of missing values:

1. Drop rows or columns that contain missing values
2. Replace by mean of the entire feature
3. Replace by mean of consecutive values