slope 0.29.0
Loading...
Searching...
No Matches
Public Member Functions | List of all members
slope::MAE Class Reference

Mean Absolute Error (MAE) scoring metric. More...

#include <score.h>

Inheritance diagram for slope::MAE:
Inheritance graph
[legend]
Collaboration diagram for slope::MAE:
Collaboration graph
[legend]

Public Member Functions

double eval (const Eigen::MatrixXd &eta, const Eigen::MatrixXd &y, const std::unique_ptr< Loss > &) const override
 
- Public Member Functions inherited from slope::MinimizeScore
bool isWorse (double a, double b) const override
 
double initValue () const override
 
- Public Member Functions inherited from slope::Score
std::function< bool(double, double)> getComparator () const
 

Additional Inherited Members

- Static Public Member Functions inherited from slope::Score
static std::unique_ptr< Score > create (const std::string &metric)
 

Detailed Description

Mean Absolute Error (MAE) scoring metric.

Computes the average absolute difference between predictions and true responses. Inherits from MinimizeScore since lower MAE values indicate better fit.

MAE = (1/n) Σ|y_i - η_i| where:

  • n is the number of samples
  • y_i is the true response for sample i
  • η_i is the predicted response for sample i

MAE is more robust to outliers compared to MSE as it uses absolute rather than squared differences.

Definition at line 189 of file score.h.

Member Function Documentation

◆ eval()

double slope::MAE::eval ( const Eigen::MatrixXd &  eta,
const Eigen::MatrixXd &  y,
const std::unique_ptr< Loss > &   
) const
overridevirtual

Evaluates the MAE between predictions and true responses.

Parameters
etaMatrix of model predictions
yMatrix of true responses
Returns
Computed MAE value

Implements slope::Score.

Definition at line 135 of file score.cpp.

The documentation for this class was generated from the following files: