liveBook · Manning

Intercept

As explained in Timeless Algorithms, The Foundational Ideas

A bias term is an additive constant that shifts a model's output independently of its input features.

An intercept is an additive constant in a model: it sets the output when feature contributions are zero. In linear regression, β₀ is the expected response at x = 0; in logistic regression, it contributes to the log-odds. In a support vector machine, the bias shifts the decision boundary without changing its orientation. Neural-network nodes also add an adjustable bias before applying an activation function. section 3.4.1 section 8.2.1 section 11.2.1

z=β0+β1x1+β2x2+⋯+βkxk

The logistic model’s linear predictor adds the intercept β₀ to the weighted feature values. section 3.3.2

When the baseline is an extrapolation

For study-hours data, the fitted regression is ŷ = 1.28 + 0.94x. The intercept predicts a score of 1.28 at zero study hours, but that interpretation warrants caution because zero lies outside the observed data range. section 3.4.3

How Timeless Algorithms, The Foundational Ideas covers it

How deep it goes
Defines the intercept’s role in regression and explains its effect on SVM boundaries and neural-network nodes, without detailing estimation for every model.
Examples
Uses a study-hours regression fit and a two-feature SVM decision function to illustrate intercept effects.
What you can do after
Evaluate/understand how an additive constant affects a predicted response, log-odds, decision boundary, or node activation.
How it connects
Connects the intercept to linear regression, logistic regression, support vector machines, and neural networks.

Also explained in

Compare how 11 books explain intercept →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Timeless Algorithms, The Foundational Ideas. The section links point to the parts of the book it draws on.