A Simple Machine Learning Example in Java

by

This is a “Hello World” example of machine learning in Java. It simply give you a taste of machine learning in Java.

Environment

Java 1.6+ and Eclipse

Step 1: Download Weka library

Download page: http://www.cs.waikato.ac.nz/ml/weka/snapshots/weka_snapshots.html

Download stable.XX.zip, unzip the file, add weka.jar to your library path of Java project in Eclipse.

Step 2: Prepare Data

Create a txt file “weather.txt” by following the following format:

@relation weather

@attribute outlook {sunny, overcast, rainy}
@attribute temperature numeric
@attribute humidity numeric
@attribute windy {TRUE, FALSE}
@attribute play {yes, no}

@data
sunny,85,85,FALSE,no
sunny,80,90,TRUE,no
overcast,83,86,FALSE,yes
rainy,70,96,FALSE,yes
rainy,68,80,FALSE,yes
rainy,65,70,TRUE,no
overcast,64,65,TRUE,yes
sunny,72,95,FALSE,no
sunny,69,70,FALSE,yes
rainy,75,80,FALSE,yes
sunny,75,70,TRUE,yes
overcast,72,90,TRUE,yes
overcast,81,75,FALSE,yes
rainy,71,91,TRUE,no

This dataset is from weka download package. It is located at “/data/weather.numeric.arff”. The file extension name is “arff”, but we can simply use “txt”.

Step 3: Training and Testing by Using Weka

This code example use a set of classifiers provided by Weka. It trains model on the given dataset and test by using 10-split cross validation. I will explain each classifier later as it is a more complicated topic.

import java.io.BufferedReader;
import java.io.FileNotFoundException;
import java.io.FileReader;
import weka.classifiers.Classifier;
import weka.classifiers.Evaluation;
import weka.classifiers.evaluation.NominalPrediction;
import weka.classifiers.rules.DecisionTable;
import weka.classifiers.rules.PART;
import weka.classifiers.trees.DecisionStump;
import weka.classifiers.trees.J48;
import weka.core.FastVector;
import weka.core.Instances;
 
public class WekaTest {
	public static BufferedReader readDataFile(String filename) {
		BufferedReader inputReader = null;
 
		try {
			inputReader = new BufferedReader(new FileReader(filename));
		} catch (FileNotFoundException ex) {
			System.err.println("File not found: " + filename);
		}
 
		return inputReader;
	}
 
	public static Evaluation classify(Classifier model,
			Instances trainingSet, Instances testingSet) throws Exception {
		Evaluation evaluation = new Evaluation(trainingSet);
 
		model.buildClassifier(trainingSet);
		evaluation.evaluateModel(model, testingSet);
 
		return evaluation;
	}
 
	public static double calculateAccuracy(FastVector predictions) {
		double correct = 0;
 
		for (int i = 0; i < predictions.size(); i++) {
			NominalPrediction np = (NominalPrediction) predictions.elementAt(i);
			if (np.predicted() == np.actual()) {
				correct++;
			}
		}
 
		return 100 * correct / predictions.size();
	}
 
	public static Instances[][] crossValidationSplit(Instances data, int numberOfFolds) {
		Instances[][] split = new Instances[2][numberOfFolds];
 
		for (int i = 0; i < numberOfFolds; i++) {
			split[0][i] = data.trainCV(numberOfFolds, i);
			split[1][i] = data.testCV(numberOfFolds, i);
		}
 
		return split;
	}
 
	public static void main(String[] args) throws Exception {
		BufferedReader datafile = readDataFile("weather.txt");
 
		Instances data = new Instances(datafile);
		data.setClassIndex(data.numAttributes() - 1);
 
		// Do 10-split cross validation
		Instances[][] split = crossValidationSplit(data, 10);
 
		// Separate split into training and testing arrays
		Instances[] trainingSplits = split[0];
		Instances[] testingSplits = split[1];
 
		// Use a set of classifiers
		Classifier[] models = { 
				new J48(), // a decision tree
				new PART(), 
				new DecisionTable(),//decision table majority classifier
				new DecisionStump() //one-level decision tree
		};
 
		// Run for each model
		for (int j = 0; j < models.length; j++) {
 
			// Collect every group of predictions for current model in a FastVector
			FastVector predictions = new FastVector();
 
			// For each training-testing split pair, train and test the classifier
			for (int i = 0; i < trainingSplits.length; i++) {
				Evaluation validation = classify(models[j], trainingSplits[i], testingSplits[i]);
 
				predictions.appendElements(validation.predictions());
 
				// Uncomment to see the summary for each training-testing pair.
				//System.out.println(models[j].toString());
			}
 
			// Calculate overall accuracy of current classifier on all splits
			double accuracy = calculateAccuracy(predictions);
 
			// Print current classifier's name and accuracy in a complicated,
			// but nice-looking way.
			System.out.println("Accuracy of " + models[j].getClass().getSimpleName() + ": "
					+ String.format("%.2f%%", accuracy)
					+ "\n---------------------------------");
		}
 
	}
}

The package view of your project should look like the following:

java-machine-learning-example

References:
1. http://www.cs.umb.edu/~ding/history/480_697_spring_2013/homework/WekaJavaAPITutorial.pdf
2. http://www.cs.ru.nl/P.Lucas/teaching/DM/weka.pdf

27 thoughts on “A Simple Machine Learning Example in Java”

  2. if the name is same then how to resolve this problem. i am using intellij idea and it is importing only that is weka.classifiers.Evaluation this. How to resolve this please help me. I need to solve this urgently.

  3. Hi how to use weka jar for association rule mining in eclipse with example code and where I can read the explaination of weka.jar function.

  4. Can you please explain what is
    Accuracy of J48 ?
    ———————————
    Accuracy of PART ?
    ———————————
    Accuracy of DecisionTable ?
    ———————————
    Accuracy of DecisionStump ?

    Thanks a lot

  5. Kevin,
    I have tried your example, with following error
    Exception in thread “main” java.lang.ClassCastException: weka.classifiers.Evaluation cannot be cast to weka.classifiers.evaluation.NominalPrediction
    at WekaTest.calculateAccuracy(WekaTest.java:46)
    at WekaTest.main(WekaTest.java:110)

  7. Output of the above programme. percentage may vary according to ur dataset.

    Accuracy of J48: 64.29%
    ———————————
    Accuracy of PART: 64.29%
    ———————————
    Accuracy of DecisionTable: 78.57%
    ———————————
    Accuracy of DecisionStump: 64.29%
    ———————————

  8. i have used another text data file.
    following is output of above prog….

    Accuracy of J48: 64.29%
    ———————————
    Accuracy of PART: 64.29%
    ———————————
    Accuracy of DecisionTable: 78.57%
    ———————————
    Accuracy of DecisionStump: 64.29%
    ———————————

  12. I guess u r importing weka.classifiers.Evaluation instead of weka.classifiers.Evaluation.
    There are 2 classes with the same name and it’s a bit confusing, but with the second one everything works fine.

  16. Hello, thanks for the tutorial, I’m having problems understanding the outputs could you explain me please.

  20. hi
    when i run this code in android emulator it unfortunately stop.We can not run it.
    please suggest some solution

  25. Hi line

    predictions.appendElements(validation.predictions()); throwing error saying validation does not have predictions method

  26. it is awesome and works well ………Can you provide me the code for support vector machine like this ……..that you I need it as soon as possible…….

Leave a Comment