What’s the interpretation of the bias term in linear models?
Classic MLLevel: IntermediateRegularization
What’s the interpretation of the bias term in linear models?
Hints
- Compare a line constrained through the origin with one that has an intercept.
- What does the prediction at zero-valued features mean, and how does centering affect it?
Answer
In a linear model, the bias term is the intercept: the predicted value when all features are zero. It shifts the fitted relationship without changing its slopes. With centered features it is the prediction at their means. This is different from statistical estimation bias or a prediction error.
Source and licence: Alexey Grigorev & contributors — Data Science Interviews · CC-BY-4.0 · answer: Quantoros editorial