MinMax Scaling Method

What is MinMax Scaling?

MinMax Scaling is a data preprocessing technique that transforms features to a specified range, typically [0, 1].

Formula

x_scaled = (x - x_min) / (x_max - x_min)

Where:
- x_scaled: Scaled value
- x: Original value
- x_min: Minimum value of the feature
- x_max: Maximum value of the feature

When to Use It

• Neural Networks: Requires bounded input values
• Image Processing: Pixel values typically scaled to [0, 1]
• Algorithms requiring bounded inputs: Some algorithms work better with bounded values

Limitations

• Sensitive to Outliers: Outliers can compress most values into a small range
• New Data: Need to know min/max of training data for new samples

Related Methods

• Standard Scaling: Uses mean and standard deviation
• Robust Scaling: Handles outliers better

Related Terms

• Feature Selection: Often done before scaling