What is an example of proper usage of the libSVM library functions? - machine-learning

Question

I need to see some example code in java so that i can figure out the proper functioning of the various methods defined in the library.Also how to pass various necessary parameters.
some of them are
svm_predict
svm_node
svm_problem etc.
I have done a lot of googling and i still haven't found something substantial. And the documentation for java is another major disappointment. please help me out!!
here is some code that i have written so far.
import java.io.BufferedInputStream;
import java.io.FileInputStream;
import libsvm.*;
import libsvm.svm_node;
import java.io.IOException;
import java.io.InputStream;
public class trial {
public static void main(String[] args) throws IOException {
svm temp = new svm();
svm_model model;
model = svm.svm_load_model("C:\\Users\\sidharth\\Desktop\\libsvm-3.18\\windows\\svm- m.model");
svm_problem prob = new svm_problem();
prob.l = trial.countLines("C:\\Users\\sidharth\\Desktop\\libsvm-3.18\\windows\\svm-ml.test");
prob.y = new double[prob.l];
int i;
for(i=0;i<prob.l;i++)
{
prob.y[i]=0.0;
}
prob.x = new svm_node[prob.l][];
temp.svm_predict(model, /*what to put here*/);
}
public static int countLines(String filename) throws IOException {
InputStream is = new BufferedInputStream(new FileInputStream(filename));
try {
byte[] c = new byte[1024];
int count = 0;
int readChars = 0;
boolean empty = true;
while ((readChars = is.read(c)) != -1) {
empty = false;
for (int i = 0; i < readChars; ++i) {
if (c[i] == '\n') {
++count;
}
}
}
return (count == 0 && !empty) ? 1 : count;
} finally {
is.close();
}
}
}
I already have a model file created and i want to predict a sample data using this model file. I have given prob.y[] a label of 0 .
Any example code that has been written by you will be of great help.
P.S. I am supposed to make an SVM based POS tagger. that is why i have tagged nlp.

Answer 1

Its pretty straightforward.
For training you could write something of the form:
svm_train training = new svm_train();
String[] options = new String[7];
options [0] = "-c";
options [1] = "1";
options [2] = "-t";
options [3] = "0"; //linear kernel
options [4] = "-v";
options [5] = "10"; //10 fold cross-validation
options [6] = your_training_filename;
training.run(options);
If you choose to save the model. Then you can retrieve it by
libsvm.svm_model model = training.getModel();
If you wish to test the model on test data, you could write something of the form:
BufferedReader input = new BufferedReader(new FileReader(test_file));
DataOutputStream output = new DataOutputStream(new BufferedOutputStream(new FileOutputStream(prediction_output_file)));
svm_predict.predict(input, output, model, 0);
Hope this helps !