use of de.unidue.ltl.evaluation.measures.regression.MeanSquaredError in project dkpro-tc by dkpro.
the class MultiRegressionUsingWekaLibsvmLiblinearTest method getMeanSquaredErrorCrossValidation.
private double getMeanSquaredErrorCrossValidation(List<File> id2outcomeFiles, String simpleName) throws Exception {
for (File f : id2outcomeFiles) {
File file = new File(f.getParentFile(), "ATTRIBUTES.txt");
Set<String> readSubTasks = readSubTasks(file);
for (String s : readSubTasks) {
File file2 = new File(f.getParentFile().getParentFile() + "/" + s, "ATTRIBUTES.txt");
if (!file2.exists()) {
continue;
}
Set<String> readSubTasks2 = readSubTasks(file2);
for (String k : readSubTasks2) {
if (k.toLowerCase().contains(simpleName.toLowerCase())) {
EvaluationData<Double> data = Tc2LtlabEvalConverter.convertRegressionModeId2Outcome(f);
MeanSquaredError mse = new MeanSquaredError(data);
return mse.getResult();
}
}
}
}
return -1;
}
use of de.unidue.ltl.evaluation.measures.regression.MeanSquaredError in project dkpro-tc by dkpro.
the class MultiRegressionUsingWekaLibsvmLiblinearTest method getMeanSquaredError.
private double getMeanSquaredError(List<File> id2outcomeFiles, String simpleName) throws Exception {
for (File f : id2outcomeFiles) {
if (f.getAbsolutePath().contains(simpleName + "TestTask-")) {
EvaluationData<Double> data = Tc2LtlabEvalConverter.convertRegressionModeId2Outcome(f);
MeanSquaredError mse = new MeanSquaredError(data);
return mse.getResult();
}
}
return -1;
}
use of de.unidue.ltl.evaluation.measures.regression.MeanSquaredError in project dkpro-tc by dkpro.
the class MetricComputationUtil method getResults.
public static Map<String, String> getResults(File id2o, String mode) throws Exception {
if (mode == null) {
throw new IllegalArgumentException("The learning mode is null");
}
Map<String, String> map = new HashMap<>();
if (mode.equals(Constants.LM_SINGLE_LABEL)) {
EvaluationData<String> data = Tc2LtlabEvalConverter.convertSingleLabelModeId2Outcome(id2o);
Accuracy<String> acc = new Accuracy<>(data);
map.put(acc.getClass().getSimpleName(), "" + acc.getResult());
} else if (mode.equals(Constants.LM_REGRESSION)) {
EvaluationData<Double> data = Tc2LtlabEvalConverter.convertRegressionModeId2Outcome(id2o);
EvaluationMeasure<?> m = new RSquared(data);
map.put(m.getClass().getSimpleName(), getExceptionFreeResult(m));
m = new PearsonCorrelation(data);
map.put(m.getClass().getSimpleName(), getExceptionFreeResult(m));
m = new SpearmanCorrelation(data);
map.put(m.getClass().getSimpleName(), getExceptionFreeResult(m));
m = new MeanSquaredError(data);
map.put(m.getClass().getSimpleName(), getExceptionFreeResult(m));
m = new MeanAbsoluteError(data);
map.put(m.getClass().getSimpleName(), getExceptionFreeResult(m));
} else if (mode.equals(Constants.LM_MULTI_LABEL)) {
EvaluationData<String> data = Tc2LtlabEvalConverter.convertMultiLabelModeId2Outcome(id2o);
EvaluationMeasure<?> m = new ExactMatchRatio<>(data);
map.put(m.getClass().getSimpleName(), getExceptionFreeResult(m));
EvaluationData<Integer> dataInt = Tc2LtlabEvalConverter.convertMultiLabelModeId2OutcomeUseInteger(id2o);
m = new HammingLoss(dataInt);
map.put(m.getClass().getSimpleName(), getExceptionFreeResult(m));
m = new MultilabelAccuracy(dataInt);
map.put(m.getClass().getSimpleName(), getExceptionFreeResult(m));
}
return map;
}
Aggregations