Ian Ma

Quick, rough guide to getting started with Weka using Java and Eclipse.

1) Make sure you’ve downloaded Weka

2) Create a new project in Eclipse. Find Java Build Path -> Libraries either during project creation or afterwards under “Package Explorer” -> RClick project -> Properties.

3) “Add External Jars…” and select the weka.jar from your download.

4) Create a class file under the “src” folder. This code is taken pretty much line for line from weka.wikispaces.

import weka.classifiers.Classifier;
import weka.classifiers.Evaluation;
import weka.classifiers.bayes.NaiveBayes;
import weka.core.Attribute;
import weka.core.FastVector;
import weka.core.Instance;
import weka.core.Instances;

public class Driver {

	public static void main(String[] args) throws Exception{

		 // Declare two numeric attributes
		 Attribute Attribute1 = new Attribute("firstNumeric");
		 Attribute Attribute2 = new Attribute("secondNumeric");

		 // Declare a nominal attribute along with its values
		 FastVector fvNominalVal = new FastVector(3);
		 fvNominalVal.addElement("blue");
		 fvNominalVal.addElement("gray");
		 fvNominalVal.addElement("black");
		 Attribute Attribute3 = new Attribute("aNominal", fvNominalVal);

		 // Declare the class attribute along with its values
		 FastVector fvClassVal = new FastVector(2);
		 fvClassVal.addElement("positive");
		 fvClassVal.addElement("negative");
		 Attribute ClassAttribute = new Attribute("theClass", fvClassVal);

		 // Declare the feature vector
		 FastVector fvWekaAttributes = new FastVector(4);
		 fvWekaAttributes.addElement(Attribute1);
		 fvWekaAttributes.addElement(Attribute2);
		 fvWekaAttributes.addElement(Attribute3);
		 fvWekaAttributes.addElement(ClassAttribute);

		 // Create an empty training set
		 Instances isTrainingSet = new Instances("Rel", fvWekaAttributes, 10);       

		 // Set class index
		 isTrainingSet.setClassIndex(3);

		 // Create the instance
		 Instance iExample = new Instance(4);
		 iExample.setValue((Attribute)fvWekaAttributes.elementAt(0), 1.0);
		 iExample.setValue((Attribute)fvWekaAttributes.elementAt(1), 0.5);
		 iExample.setValue((Attribute)fvWekaAttributes.elementAt(2), "gray");
		 iExample.setValue((Attribute)fvWekaAttributes.elementAt(3), "positive");

		 // add the instance
		 isTrainingSet.add(iExample);
		 Classifier cModel = (Classifier)new NaiveBayes();
		 cModel.buildClassifier(isTrainingSet);

		 // Test the model
		 Evaluation eTest = new Evaluation(isTrainingSet);
		 eTest.evaluateModel(cModel, isTrainingSet);

		 // Print the result à la Weka explorer:
		 String strSummary = eTest.toSummaryString();
		 System.out.println(strSummary);

		 // Get the confusion matrix
		 double[][] cmMatrix = eTest.confusionMatrix();
		 for(int row_i=0; row_i<cmMatrix.length; row_i++){
			 for(int col_i=0; col_i<cmMatrix.length; col_i++){
				 System.out.print(cmMatrix[row_i][col_i]);
				 System.out.print("|");
			 }
			 System.out.println();
		 }
	}
}

I’m building a predictive model that’s time-series related. I wanted to visualize the instances I predicted incorrectly on a time-series plot. The first step, however, is to get it into a format that R is happy with. Here’s how to add a “predicted” column to your training file.

BASIC (only need existing columns):

In the Weka explorer, under the classify tab, click “More Options”. Make sure the “Store predictions for visualization” is checked.

Click “Start” to build and run the model.

When finished, right-click the model name from the Result List. Click on “Visualize classifier errors.”

Click “Save” in that new window and the outputted file will have the new predicted column.

To convert the resulting ARFF file to CSV, do “java weka.core.converters.CSVSaver -i your.arff -o your.csv”

ADVANCED (need excluded columns):

Say you have extra columns for debugging that you need to exclude before you use it for training — think instance IDs or date markers. Here’s how you would do that.

In the Explorer GUI, go to the classify tab

Choose “FilteredClassifier” under the “meta” folder

Go inside the FilteredClassifier options and choose your base classifier (J48)

In “filter” option, remove “AllFilter” and add “Unsupervised -> Attribute -> Remove”.

In the “Remove” option, choose the attribute index that you want to remove. Then click Add.

You’re now ready to run your model. Follow the latter steps above in “BASIC” to go through visualization and save the arff with the predicted column.