LogitLink<T>
Models & Types · AiDotNet.LinkFunctions
Logit link function: g(μ) = log(μ/(1-μ)).
Implements: ILinkFunction<T>
For Beginners
Use this for binary classification (logistic regression). It ensures your predictions are valid probabilities between 0 and 1.
The logit function is the log-odds:
- logit(0.5) = 0 (50-50 odds)
- logit(0.9) = 2.2 (9:1 odds)
- logit(0.1) = -2.2 (1:9 odds)
The inverse logit (sigmoid) converts back:
- sigmoid(0) = 0.5
- sigmoid(2.2) ≈ 0.9
- sigmoid(-2.2) ≈ 0.1
How It Works
The logit link is the canonical link for the Binomial distribution. It maps probabilities from (0,1) to (-∞,∞).
Constructors
| Constructor | Summary |
|---|---|
LogitLink() |
Properties
| Property | Summary |
|---|---|
bool IsCanonical { get; } |
|
string Name { get; } |
Methods
| Method | Summary |
|---|---|
T InverseLink(T eta) |
|
T InverseLinkDerivative(T eta) |
|
T Link(T mu) |
|
T LinkDerivative(T mu) |
|
T Variance(T mu) |