package jsat.regression.evaluation;

import java.io.Serializable;

/**
 * This interface defines the contract for evaluating or "scoring" the results 
 * on a regression problem. 
*
* All regression scores must override the {@link #equals(java.lang.Object)} * and {@link #hashCode() } methods. If a score has parameters, different * objects with different parameters must not be equal. However, different * objects with the same parameters must be equal even if their internal * states are different * * @author Edward Raff */ public interface RegressionScore extends Serializable { public void prepare(); /** * Adds the given result to the score * @param prediction the prediction for the data point * @param trueValue the true value for the data point * @param weight the weigh to assign to the data point */ public void addResult(double prediction, double trueValue, double weight); /** * The score contained in this object is augmented with the results * already accumulated in the {@code other} object. This does not result in * an averaging, but alters the current object to have the same score it * would have had if all the results were originally inserted into this * object.
*
* This method is only required to work if {@code other} if of the same * class as {@code this} object. * * @param other the object to add the results from */ public void addResults(RegressionScore other); /** * Computes the score for the results that have been enrolled via * {@link #addResult(double, double, double) } * * @return the score for the current results */ public double getScore(); /** * Returns {@code true} if a lower score is better, or {@code false} if a * higher score is better * @return {@code true} if a lower score is better */ public boolean lowerIsBetter(); @Override public boolean equals(Object obj); @Override public int hashCode(); public RegressionScore clone(); /** * Returns the name to present for this score * @return the score name */ public String getName(); }

Ads help maintain this website.