Explanation:
Feature scaling techniques, like Min-Max scaling, are crucial before model training in Spark ML workflows. This step is essential for algorithms sensitive to feature scales, such as those based on distance metrics or optimization algorithms like Gradient Descent. Scaling ensures all features contribute equally to the learning process by standardizing their ranges, for instance, between 0 and 1 with Min-Max scaling. This preprocessing step helps in preventing any single feature from disproportionately influencing the model's performance.
No comments yet.
In the context of Spark ML workflows, at which stage should feature scaling techniques such as Min-Max scaling be applied?
A
Feature scaling is not necessary in Spark ML
B
After model training
C
Before data preprocessing
D
Before model training