Workflow Automation Templates
A library of ready-to-use workflow templates to accelerate your data journey
Feature Scaling
Standardize and rescale numerical data
Overview
This workflow scales numerical features to ensure consistent data ranges for better model performance and convergence. It applies both Standard Scaler and MinMax Scaler techniques to normalize features and stabilize learning algorithms.
Details
The process begins by reading the housing dataset and assembling features into a single vector using Vector Assembler. The Standard Scaler node normalizes each feature to have zero mean and unit variance, while the MinMax Scaler rescales values to a defined range (commonly 0 to 1).
The transformed data is displayed using Print N Rows, allowing users to compare both scaling techniques. This workflow improves model training by ensuring all features contribute proportionally, preventing scale-dominated biases.