Interface Gaussian
public interface Gaussian
The response variable is of Gaussian (normal) distribution.
This family is used for ordinary linear regression and Gaussian GAMs.
The canonical link is the identity function, giving the standard
linear-Gaussian model.
The Gaussian family supports three link functions:
identity()– the canonical identity linkg(mu) = mu.log()– the log linkg(mu) = log(mu), useful when the mean must be positive.inverse()– the inverse linkg(mu) = 1/mu.
-
Method Summary
-
Method Details
-
identity
Identity link function (canonical link for Gaussian).g(mu) = mu,g^{-1}(eta) = eta.- Returns:
- the identity link model.
-
log
Log link function for Gaussian.g(mu) = log(mu),g^{-1}(eta) = exp(eta). Useful when the mean must be positive.- Returns:
- the log link model.
-
inverse
Inverse link function for Gaussian.g(mu) = 1/mu,g^{-1}(eta) = 1/eta.- Returns:
- the inverse link model.
-