Java Code for Spam email/sms Classification using Weka Machine learning

WekaClassifier.java

			import weka.classifiers.Evaluation;
			import weka.classifiers.bayes.NaiveBayes;
			import weka.classifiers.bayes.NaiveBayesMultinomial;
			import weka.classifiers.meta.FilteredClassifier;
			import weka.classifiers.Evaluation;

			import weka.core.Instances;
			import weka.core.Instance;
			import weka.core.converters.ConverterUtils.DataSource;
			import weka.core.Attribute;
			import weka.core.DenseInstance;
			import weka.core.converters.ArffSaver;
			import weka.classifiers.meta.FilteredClassifier;

			import weka.filters.unsupervised.attribute.StringToWordVector;

			import java.io.File;
			import java.io.BufferedReader;
			import java.io.FileReader;
			import java.io.IOException;

			import java.util.List;
			import java.util.ArrayList;

			public class WekaClassifier {

				private FilteredClassifier classifier;
				private Instances trainData;
				private Instances testData;
				private ArrayList<Attribute> fvWekaAttributes;

				WekaClassifier(){
					classifier = new FilteredClassifier();
								// Declare text attribute
					 Attribute attribute_text = new Attribute("text",(List<String>) null);

					 // Declare the label attribute along with its values
					 ArrayList<String> classAttributeValues = new ArrayList<String>();
					 classAttributeValues.add("spam");
					 classAttributeValues.add("ham");
					 Attribute classAttribute = new Attribute("label", classAttributeValues);

					 // Declare the feature vector
					 fvWekaAttributes = new ArrayList<Attribute>();
					 fvWekaAttributes.add(classAttribute);
					 fvWekaAttributes.add(attribute_text);

				}
				public Instances load (String filename)  throws IOException
				{

					// http://geekswithblogs.net/razan/archive/2011/11/08/creating-a-simple-sparse-arff-file.aspx
					// http://weka.wikispaces.com/Programmatic+Use

					// // Declare text attribute
					//  Attribute attribute_text = new Attribute("text",(List<String>) null);

					//  // Declare the label attribute along with its values
					//  ArrayList<String> classAttributeValues = new ArrayList<String>();
					//  classAttributeValues.add("spam");
					//  classAttributeValues.add("ham");
					//  Attribute classAttribute = new Attribute("label", classAttributeValues);

					//  // Declare the feature vector
					//  ArrayList<Attribute> fvWekaAttributes = new ArrayList<Attribute>();
					//  fvWekaAttributes.add(classAttribute);
					//  fvWekaAttributes.add(attribute_text);

					 /* 
					    Create an empty training set
						name the relation “Rel”.
						set intial capacity of 10*
					*/	
					 Instances dataset = new Instances("Rel", fvWekaAttributes, 10);

					 // Set class index
					 dataset.setClassIndex(0);

					 // read text file, parse data and add to instance
					try(BufferedReader br = new BufferedReader(new FileReader(filename))) {
					    for(String line; (line = br.readLine()) != null; ) {
					    	try{

					    		// split at first occurance of n no. of words
					        	String parts[] = line.split("\\s+",2);

					        	 // basic validation
					        	if (!parts[0].isEmpty() && !parts[1].isEmpty()){

					        	  DenseInstance row = new DenseInstance(2);
								  row.setValue(fvWekaAttributes.get(0), parts[0]);
								  row.setValue(fvWekaAttributes.get(1), parts[1]);	

								  // add row to instances
								  dataset.add(row);
					        	}
								// 
					    	}
					    	catch (ArrayIndexOutOfBoundsException e){
								System.out.println("invalid row");
							}

					    }

					}
					catch (IOException e){
						e.printStackTrace();
					}
					 return dataset;

				}

				public void prepare() throws Exception{
					trainData = load("data/train.txt");
					testData = load("data/test.txt");
				}

				public void transform(){

					// create the filter and set the attribute to be transformed from text into a feature vector (the last one)
					StringToWordVector filter = new StringToWordVector();
					filter.setAttributeIndices("last"); 

					classifier.setFilter(filter); 
					classifier.setClassifier(new NaiveBayes());

				}
				public void fit() throws Exception{
					classifier.buildClassifier(trainData);
				}

				public String classify(String text) throws Exception  {

						Instances newDataset = new Instances("testdata", fvWekaAttributes, 1);
						newDataset.setClassIndex(0);
					
						DenseInstance newinstance = new DenseInstance(2);
						newinstance.setDataset(newDataset); 

						newinstance.setValue(fvWekaAttributes.get(1), text);

						double pred = classifier.classifyInstance(newinstance);

						System.out.println("===== Classified instance =====");
						System.out.println("Class predicted: " + trainData.classAttribute().value((int) pred));
						return trainData.classAttribute().value((int) pred);

					// try {
					// 	DenseInstance instance = new DenseInstance(2);
					// 	instance.setValue(new Attribute("text",(List<String>) null), text);
					// 	double pred = classifier.classifyInstance(instance);
					// 	System.out.println("===== Classified instance =====");
					// 	System.out.println("Class predicted: " + trainData.classAttribute().value((int) pred));
					// 	return trainData.classAttribute().value((int) pred);
					// }
					// catch (Exception e) {
					// 	System.out.println("Problem found when classifying the text");
					// }	
					// return "";	
				}

				public String evaluate() throws Exception{
					Evaluation eval = new Evaluation(testData);
					eval.evaluateModel(classifier, testData);
					System.out.println(eval.toSummaryString());
					return eval.toSummaryString();
				}


				public void saveArff(Instances dataset,String filename)   throws IOException{
					try
					{
						// initialize 
					   ArffSaver arffSaverInstance = new ArffSaver(); 
				       arffSaverInstance.setInstances(dataset); 
				       arffSaverInstance.setFile(new File(filename)); 
				       arffSaverInstance.writeBatch();	
					}
					catch (IOException e){
						e.printStackTrace();
					}
				}

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

					WekaClassifier wt = new WekaClassifier();
					wt.prepare();
					wt.transform();
					wt.fit();
					wt.evaluate();
					wt.classify("goldviking (29/M) is inviting you to be his friend. Reply YES-762 or NO-762 See him: www.SMS.ac/u/goldviking STOP? Send STOP FRND to 62468");

					// Instances trainData = wt.load("data/train.txt");
					// System.out.println(trainData);

					 
					// // create the filter and set the attribute to be transformed from text into a feature vector (the last one)
					// StringToWordVector filter = new StringToWordVector();
					// filter.setAttributeIndices("last"); 

					// FilteredClassifier classifier = new FilteredClassifier(); 
					// classifier.setFilter(filter); 
					// classifier.setClassifier(new NaiveBayes());

					// classifier.buildClassifier(trainData);

					// System.out.println(WekaTransformer.classify("I said its okay. Sorry"));

					/*
					  Now that we create and trained a classifier, let’s test it.
					  To do so, we need an evaluation module (weka.classifiers.Evaluation) to which we feed our testing set
					*/

					// Instances testData = wt.transform("data/test.txt");
					// evaluation set
					// Evaluation eval = new Evaluation(testData);
					// eval.evaluateModel(classifier, testData);
					// System.out.println(eval.toSummaryString());






				}
			}

Hello good work

Nice Work!