/**
* Generating the header of the output file @ param p PrintStream
*
* @return Nothing
*/
private void CopyHeaderTest(PrintStream p) {
// Header of the output file
p.println("@relation " + Attributes.getRelationName());
p.print(Attributes.getInputAttributesHeader());
p.print(Attributes.getOutputAttributesHeader());
p.println(Attributes.getInputHeader());
p.println(Attributes.getOutputHeader());
p.println("@data");
}
/**
* It copies the header of the dataset
*
* @return String A string containing all the data-set information
*/
public String copyHeader() {
String p = new String("");
p = "@relation " + Attributes.getRelationName() + "\n";
p += Attributes.getInputAttributesHeader();
p += Attributes.getOutputAttributesHeader();
p += Attributes.getInputHeader() + "\n";
p += Attributes.getOutputHeader() + "\n";
p += "@data\n";
return p;
}
/**
* Save data in output file
*
* @param file_name File name
* @param data Data to be saved
* @param n No of patterns
* @param problem Type of problem (CLASSIFICATION | REGRESSION)
* @throws IOException
*/
public void SaveOutputFile(
String file_name, double data[][], int n, String problem, double[] a, double[] b) {
String line;
double outputs[] = new double[Noutputs];
try {
// Result file
FileOutputStream file = new FileOutputStream(file_name);
BufferedWriter f = new BufferedWriter(new OutputStreamWriter(file));
// File header
f.write("@relation " + Attributes.getRelationName() + "\n");
f.write(Attributes.getInputAttributesHeader());
f.write(Attributes.getOutputAttributesHeader());
f.write(Attributes.getInputHeader() + "\n");
f.write(Attributes.getOutputHeader() + "\n");
f.write("@data\n");
// For all patterns
for (int i = 0; i < n; i++) {
// Classification
if (problem.compareToIgnoreCase("Classification") == 0) {
// Obtain class
int Class = 0;
for (int j = 1; j < Noutputs; j++) {
if (data[i][Class + Ninputs] < data[i][j + Ninputs]) {
Class = j;
}
}
/*
f.write(Integer.toString(Class) + " ");
f.write(Integer.toString(EnsembleGetClassOfPattern(data[i])));
f.newLine();
*/
f.write(Attributes.getOutputAttributes()[0].getNominalValue(Class) + " ");
f.write(
Attributes.getOutputAttributes()[0].getNominalValue(
EnsembleGetClassOfPattern(data[i])));
f.newLine();
f.flush();
}
// Regression
else {
if (a != null && b != null) {
for (int j = 0; j < Noutputs; j++) {
f.write(Double.toString((data[i][Ninputs + j] - b[j]) / a[j]) + " ");
}
EnsembleOutput(data[i], outputs);
for (int j = 0; j < Noutputs; j++) {
f.write(Double.toString((outputs[j] - b[j]) / a[j]) + " ");
}
f.newLine();
} else {
for (int j = 0; j < Noutputs; j++) {
f.write(Double.toString(data[i][Ninputs + j]) + " ");
}
EnsembleOutput(data[i], outputs);
for (int j = 0; j < Noutputs; j++) {
f.write(Double.toString(outputs[j]) + " ");
}
f.newLine();
}
}
}
f.close();
file.close();
} catch (FileNotFoundException e) {
System.err.println("Training file does not exist");
System.exit(1);
} catch (IOException e) {
e.printStackTrace();
System.exit(-1);
}
}
