edu.uci.jforestsx.config.TrainingConfig Maven / Gradle / Ivy
/**
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package edu.uci.jforestsx.config;
import java.util.Map.Entry;
import edu.uci.jforestsx.util.ConfigHolder;
import edu.uci.jforestsx.util.IOUtils;
/**
* @author Yasser Ganjisaffar
*/
public class TrainingConfig extends ComponentConfig {
/**
* Training file name
*/
public String trainFilename = null;
public final static String TRAIN_FILENAME = "input.train";
/**
* Validation file name
*/
public String validFilename = null;
public final static String VALID_FILENAME = "input.valid";
/**
* If feature names are needed and they are not stored in the training file,
* they can be loaded from this external file. Each line in the file contains
* one feature name.
*/
public String featureNamesFilename = null;
private final static String FEATURENAMES_FILENAME = "input.train.feature-names-file";
/**
* The name of the algorithm to be used for training.
* For example, the value of "Bagging-RegressionTree" means that we want Bagging wrapped
* around Regression trees.
*/
public String learningAlgorithm = null;
private final static String LEARNING_ALGORITHM = "learning.algorithm";
/**
* The name of the evaluation metric to be used during training.
* Default is AUC.
* Other currently implemented metrics include RMSE and BalancedYoundenIndex.
*/
public String evaluationMetric = "AUC";
private final static String LEARNING_EVALUATION_METRIC = "learning.evaluation-metric";
/**
* If this parameter is set to a value less than 1.0, only a fraction of
* the training data will be used for training.
*/
public double trainFraction = 1.0;
private final static String TRAIN_FRACTION = "input.train-fraction";
/**
* If this parameter is set to a value less than 1.0, only a fraction of
* the validation data will be used for validation.
*/
public double validFraction = 1.0;
private final static String VALID_FRACTION = "input.valid-fraction";
/**
* If for training only a fraction of data is used and this parameter is set to
* true, then for validation we will use the data in training input file which
* is left out of training.
*/
public boolean validOutOfTrain = false;
private final static String VALID_OUT_OF_TRAIN = "input.valid.out-of-train";
/**
* Number of threads to use. By default this is set to the number of processors
* on the machine. However, for debugging, sometimes it is needed to set it to 1.
*/
public int numThreads = Runtime.getRuntime().availableProcessors();
private final static String NUM_THREADS = "params.num-threads";
/**
* The random seed to be used for training.
*/
public int randomSeed = 1;
private final static String RANDOM_SEED = "params.random-seed";
/**
* If this parameter is set to true, the performance on validation data is
* printed during the iterations of training.
*/
public boolean printIntermediateValidMeasurements = false;
private final static String PRINT_INTERMEDIATE_VALID_MEASUREMENTS = "params.print-intermediate-valid-measurements";
public void init(ConfigHolder config) {
for (Entry entry : config.getEntries()) {
String key = ((String) entry.getKey()).toLowerCase();
String value = (String) entry.getValue();
if (key.equals(TRAIN_FILENAME)) {
trainFilename = value;
} else if (key.equals(VALID_FILENAME)) {
validFilename = value;
} else if (key.equals(FEATURENAMES_FILENAME)) {
featureNamesFilename = value;
} else if (key.equals(LEARNING_ALGORITHM)) {
learningAlgorithm = value;
} else if (key.equals(LEARNING_EVALUATION_METRIC)) {
evaluationMetric = value;
} else if (key.equals(TRAIN_FRACTION)) {
trainFraction = Double.parseDouble(value);
} else if (key.equals(VALID_FRACTION)) {
validFraction = Double.parseDouble(value);
} else if (key.equals(NUM_THREADS)) {
numThreads = Integer.parseInt(value);
} else if (key.equals(RANDOM_SEED)) {
randomSeed = Integer.parseInt(value);
} else if (key.equals(VALID_OUT_OF_TRAIN)) {
validOutOfTrain = Boolean.parseBoolean(value);
} else if (key.equals(PRINT_INTERMEDIATE_VALID_MEASUREMENTS)) {
printIntermediateValidMeasurements = Boolean.parseBoolean(value);
}
}
}
public boolean validate(IOUtils ioUtils) {
if (trainFilename == null) {
errorMessage = "Training file is not set";
return false;
}
if (!ioUtils.exists(trainFilename)) {
errorMessage = "Training file: " + trainFilename + " does not exists.";
return false;
}
if (validFilename != null && !ioUtils.exists(validFilename)) {
errorMessage = "Validation file: " + validFilename + " does not exists.";
return false;
}
if (learningAlgorithm == null) {
errorMessage = "Learning algorithm is not specified.";
return false;
}
return true;
}
@Override
public String toString() {
StringBuilder sb = new StringBuilder();
sb.append(TRAIN_FILENAME + ": " + trainFilename + "\n");
sb.append(VALID_FILENAME + ": " + validFilename + "\n");
sb.append(FEATURENAMES_FILENAME + ": " + featureNamesFilename + "\n");
sb.append(LEARNING_ALGORITHM + ": " + learningAlgorithm + "\n");
sb.append(TRAIN_FRACTION + ": " + trainFraction + "\n");
sb.append(VALID_FRACTION + ": " + validFraction + "\n");
sb.append(NUM_THREADS + ": " + numThreads + "\n");
sb.append(RANDOM_SEED + ": " + randomSeed + "\n");
sb.append(PRINT_INTERMEDIATE_VALID_MEASUREMENTS + ": " + printIntermediateValidMeasurements + "\n");
return sb.toString();
}
}