smile.validation.ClassificationValidation Maven / Gradle / Ivy
The newest version!
/*
* Copyright (c) 2010-2021 Haifeng Li. All rights reserved.
*
* Smile is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* Smile is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with Smile. If not, see the model type.
*
* @author Haifeng
*/
public class ClassificationValidation implements Serializable {
private static final long serialVersionUID = 2L;
/** The model. */
public final M model;
/** The true class labels of validation data. */
public final int[] truth;
/** The model prediction. */
public final int[] prediction;
/** The posteriori probability of prediction if the model is a soft classifier. */
public final double[][] posteriori;
/** The confusion matrix. */
public final ConfusionMatrix confusion;
/** The classification metrics. */
public final ClassificationMetrics metrics;
/**
* Constructor.
* @param model the model.
* @param fitTime the time in milliseconds of fitting the model.
* @param scoreTime the time in milliseconds of scoring the validation data.
* @param truth the ground truth.
* @param prediction the predictions.
*/
public ClassificationValidation(M model, double fitTime, double scoreTime, int[] truth, int[] prediction) {
this.model = model;
this.truth = truth;
this.prediction = prediction;
this.posteriori = null;
this.confusion = ConfusionMatrix.of(truth, prediction);
this.metrics = ClassificationMetrics.of(fitTime, scoreTime, truth, prediction);
}
/**
* Constructor of soft classifier validation.
* @param model the model.
* @param fitTime the time in milliseconds of fitting the model.
* @param scoreTime the time in milliseconds of scoring the validation data.
* @param truth the ground truth.
* @param prediction the predictions.
* @param posteriori the posteriori probabilities of predictions.
*/
public ClassificationValidation(M model, double fitTime, double scoreTime, int[] truth, int[] prediction, double[][] posteriori) {
this.model = model;
this.truth = truth;
this.prediction = prediction;
this.posteriori = posteriori;
this.confusion = ConfusionMatrix.of(truth, prediction);
this.metrics = ClassificationMetrics.of(fitTime, scoreTime, truth, prediction, posteriori);
}
@Override
public String toString() {
return metrics.toString();
}
/**
* Trains and validates a model on a train/validation split.
* @param x the training data.
* @param y the class labels of training data.
* @param testx the validation data.
* @param testy the class labels of validation data.
* @param trainer the lambda to train the model.
* @param the data type of samples.
* @param the model type.
* @return the validation results.
*/
public static > ClassificationValidation of(T[] x, int[] y, T[] testx, int[] testy, BiFunction trainer) {
long start = System.nanoTime();
M model = trainer.apply(x, y);
double fitTime = (System.nanoTime() - start) / 1E6;
start = System.nanoTime();
if (model.soft()) {
int k = model.numClasses();
double[][] posteriori = new double[testx.length][k];
int[] prediction = model.predict(testx, posteriori);
double scoreTime = (System.nanoTime() - start) / 1E6;
return new ClassificationValidation<>(model, fitTime, scoreTime, testy, prediction, posteriori);
} else {
int[] prediction = model.predict(testx);
double scoreTime = (System.nanoTime() - start) / 1E6;
return new ClassificationValidation<>(model, fitTime, scoreTime, testy, prediction);
}
}
/**
* Trains and validates a model on multiple train/validation split.
* @param bags the data splits.
* @param x the training data.
* @param y the class labels.
* @param trainer the lambda to train the model.
* @param the data type of samples.
* @param the model type.
* @return the validation results.
*/
public static > ClassificationValidations of(Bag[] bags, T[] x, int[] y, BiFunction trainer) {
List> rounds = new ArrayList<>(bags.length);
for (Bag bag : bags) {
T[] trainx = MathEx.slice(x, bag.samples);
int[] trainy = MathEx.slice(y, bag.samples);
T[] testx = MathEx.slice(x, bag.oob);
int[] testy = MathEx.slice(y, bag.oob);
rounds.add(of(trainx, trainy, testx, testy, trainer));
}
return new ClassificationValidations<>(rounds);
}
/**
* Trains and validates a model on a train/validation split.
* @param formula the model formula.
* @param train the training data.
* @param test the validation data.
* @param trainer the lambda to train the model.
* @param the model type.
* @return the validation results.
*/
public static ClassificationValidation of(Formula formula, DataFrame train, DataFrame test, BiFunction trainer) {
int[] y = formula.y(train).toIntArray();
int[] testy = formula.y(test).toIntArray();
long start = System.nanoTime();
M model = trainer.apply(formula, train);
double fitTime = (System.nanoTime() - start) / 1E6;
int n = test.nrow();
int[] prediction = new int[n];
if (model.soft()) {
int k = model.numClasses();
double[][] posteriori = new double[n][k];
start = System.nanoTime();
for (int i = 0; i < n; i++) {
prediction[i] = model.predict(test.get(i), posteriori[i]);
}
double scoreTime = (System.nanoTime() - start) / 1E6;
return new ClassificationValidation<>(model, fitTime, scoreTime, testy, prediction, posteriori);
} else {
start = System.nanoTime();
for (int i = 0; i < n; i++) {
prediction[i] = model.predict(test.get(i));
}
double scoreTime = (System.nanoTime() - start) / 1E6;
return new ClassificationValidation<>(model, fitTime, scoreTime, testy, prediction);
}
}
/**
* Trains and validates a model on multiple train/validation split.
* @param bags the data splits.
* @param formula the model formula.
* @param data the data.
* @param trainer the lambda to train the model.
* @param the model type.
* @return the validation results.
*/
public static ClassificationValidations of(Bag[] bags, Formula formula, DataFrame data, BiFunction trainer) {
List> rounds = new ArrayList<>(bags.length);
for (Bag bag : bags) {
rounds.add(of(formula, data.of(bag.samples), data.of(bag.oob), trainer));
}
return new ClassificationValidations<>(rounds);
}
}