public static void f(String[] args) throws IOException, ClassifierException {
Vector<LabeledSample> samples = readDataFile(new File(args[0]));
printSamples(samples);
LogisticRegressionClassifier classifier =
new LogisticRegressionClassifier(LogisticRegressionClassifier.DEFAULT_LEARNING_RATE, 1.0);
classifier.train(samples);
System.out.println("Trained with " + classifier.getNumberOfIterations() + " iterations.");
System.out.println(classifier.getFeatures());
int all = 0;
int good = 0;
for (LabeledSample sample : samples) {
double classification = classifier.classify(sample.getFeatures());
if (((classification < 0.5) && (sample.getLabel() == false))
|| ((classification >= 0.5) && (sample.getLabel() == true))) good++;
all++;
}
System.out.println("correct: " + good + ". All=" + all);
System.out.println("accuracy: " + String.format("%4.4f", ((double) good) / ((double) all)));
System.out.println(StringUtil.generateStringOfCharacter('-', 50));
System.out.println(classifier.descriptionOfTraining());
System.out.println(StringUtil.generateStringOfCharacter('-', 50));
Set<Integer> dontChange = new LinkedHashSet<Integer>();
dontChange.add(2);
classifier.setToZeroNegativeParametersBut(dontChange);
System.out.println(classifier.descriptionOfTraining());
}
public static void printSamples(Vector<LabeledSample> samples) {
for (LabeledSample sample : samples) {
System.out.print(sample.getLabel() + " ");
for (Integer index : sample.getFeatures().keySet()) {
System.out.print(
index.toString()
+ ":"
+ String.format("%3.3f", sample.getFeatures().get(index))
+ ", ");
}
System.out.println();
}
}