ml::LinearRegression::RegularisedRegressionResult Struct Reference

Result of a multivariate regularised regression with intercept. More...

#include <LinearRegression.hpp>

Inheritance diagram for ml::LinearRegression::RegularisedRegressionResult:

Collaboration diagram for ml::LinearRegression::RegularisedRegressionResult:

Public Member Functions
Eigen::VectorXd	predict (Eigen::Ref< const Eigen::MatrixXd > X) const
	Predicts Y given X. More...
double	predict_single (Eigen::Ref< const Eigen::VectorXd > x) const
	Predicts Y given X. More...
Public Member Functions inherited from ml::LinearRegression::Result
double	var_y () const
	Estimated variance of observations Y, equal to rss / dof.
double	r2 () const
	R2 coefficient. More...
double	adjusted_r2 () const
	Adjusted R2 coefficient. More...

Public Attributes
Eigen::VectorXd	beta
double	effective_dof
Public Attributes inherited from ml::LinearRegression::Result
unsigned int	n
unsigned int	dof
double	rss
double	tss

Detailed Description

Result of a multivariate regularised regression with intercept.

Regularisation is applied to everything except the intercept, which is the last coefficient in beta.

Contrary to MultivariateOLSResult, it does not assume that inputs X contain a row of 1s.

var_y is calculated using dof as the denominator.

Member Function Documentation

predict()

Eigen::VectorXd ml::LinearRegression::RegularisedRegressionResult::predict

(

Eigen::Ref< const Eigen::MatrixXd >

X

)

const

Predicts Y given X.

Parameters

X	Matrix of independent variables with data points in columns.

Returns

Vector of predicted Y(X) with size X.cols().

Exceptions

std::invalid_argument

If X.rows() + 1 != beta.size().

predict_single()

double ml::LinearRegression::RegularisedRegressionResult::predict_single

(

Eigen::Ref< const Eigen::VectorXd >

x

)

const

Predicts Y given X.

Parameters

x	Vector of independent variables.

Returns

Predicted Y(X).

Exceptions

std::invalid_argument

If X.size() + 1 != beta.size().

Member Data Documentation

beta

Eigen::VectorXd ml::LinearRegression::RegularisedRegressionResult::beta

Fitted coefficients of the model \(\hat{y} = \vec{\beta'} \cdot \vec{x} + \beta_0 \), concatenated as \( (\vec{\beta'}, \beta_0) \).

effective_dof

double ml::LinearRegression::RegularisedRegressionResult::effective_dof

Effective number of residual degrees of freedom \( N - \mathrm{tr} [ X^T (X X^T + \lambda I)^{-1} X ] - 1 \).

---

The documentation for this struct was generated from the following file:

• ML/LinearRegression.hpp