Machine learning for language toolkit

Tutorial slides / video
Quick Start
Importing Data
Data Transformations
Sequence Tagging
Topic Modeling

View the Project on GitHub mimno/Mallet

Topic Modeling for Java Developers

In this example, I import data from a file, train a topic model, and analyze the topic assignments of the first instance. I then create a new instance, which is made up of the words from topic 0, and infer a topic distribution for that instance.

The input file contains one document per line. Each line has three fields, separated by commas. This is a standard Mallet format. For more information, see the importing data guide. The first field is a name for the document. The second field could contain a document label, as in a classification task, but for this example we won’t use that field. It is therefore set to a meaningless placeholder value. The third field contains the full text of the document, with no newline characters.

The following example is in the cc.mallet.examples package. Annotations are included in comments. You can run this code using the command

bin/mallet run cc.mallet.examples.TopicModel [filename]

Java source code example

package cc.mallet.examples;

import cc.mallet.util.*;
import cc.mallet.types.*;
import cc.mallet.pipe.*;
import cc.mallet.pipe.iterator.*;
import cc.mallet.topics.*;

import java.util.*;
import java.util.regex.*;

public class TopicModel {

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

        // Begin by importing documents from text to feature sequences
        ArrayList<Pipe> pipeList = new ArrayList<Pipe>();

        // Pipes: lowercase, tokenize, remove stopwords, map to features
        pipeList.add( new CharSequenceLowercase() );
        pipeList.add( new CharSequence2TokenSequence(Pattern.compile("\\p{L}[\\p{L}\\p{P}]+\\p{L}")) );
        pipeList.add( new TokenSequenceRemoveStopwords(new File("stoplists/en.txt"), "UTF-8", false, false, false) );
        pipeList.add( new TokenSequence2FeatureSequence() );

        InstanceList instances = new InstanceList (new SerialPipes(pipeList));

        Reader fileReader = new InputStreamReader(new FileInputStream(new File(args[0])), "UTF-8");
        instances.addThruPipe(new CsvIterator (fileReader, Pattern.compile("^(\\S*)[\\s,]*(\\S*)[\\s,]*(.*)$"),
                                               3, 2, 1)); // data, label, name fields

        // Create a model with 100 topics, alpha_t = 0.01, beta_w = 0.01
        //  Note that the first parameter is passed as the sum over topics, while
        //  the second is the parameter for a single dimension of the Dirichlet prior.
        int numTopics = 100;
        ParallelTopicModel model = new ParallelTopicModel(numTopics, 1.0, 0.01);


        // Use two parallel samplers, which each look at one half the corpus and combine
        //  statistics after every iteration.

        // Run the model for 50 iterations and stop (this is for testing only, 
        //  for real applications, use 1000 to 2000 iterations)

        // Show the words and topics in the first instance

        // The data alphabet maps word IDs to strings
        Alphabet dataAlphabet = instances.getDataAlphabet();

        FeatureSequence tokens = (FeatureSequence) model.getData().get(0).instance.getData();
        LabelSequence topics = model.getData().get(0).topicSequence;

        Formatter out = new Formatter(new StringBuilder(), Locale.US);
        for (int position = 0; position < tokens.getLength(); position++) {
            out.format("%s-%d ", dataAlphabet.lookupObject(tokens.getIndexAtPosition(position)), topics.getIndexAtPosition(position));

        // Estimate the topic distribution of the first instance, 
        //  given the current Gibbs state.
        double[] topicDistribution = model.getTopicProbabilities(0);

        // Get an array of sorted sets of word ID/count pairs
        ArrayList<TreeSet<IDSorter>> topicSortedWords = model.getSortedWords();

        // Show top 5 words in topics with proportions for the first document
        for (int topic = 0; topic < numTopics; topic++) {
            Iterator<IDSorter> iterator = topicSortedWords.get(topic).iterator();

            out = new Formatter(new StringBuilder(), Locale.US);
            out.format("%d\t%.3f\t", topic, topicDistribution[topic]);
            int rank = 0;
            while (iterator.hasNext() && rank < 5) {
                IDSorter idCountPair =;
                out.format("%s (%.0f) ", dataAlphabet.lookupObject(idCountPair.getID()), idCountPair.getWeight());

        // Create a new instance with high probability of topic 0
        StringBuilder topicZeroText = new StringBuilder();
        Iterator<IDSorter> iterator = topicSortedWords.get(0).iterator();

        int rank = 0;
        while (iterator.hasNext() && rank < 5) {
            IDSorter idCountPair =;
            topicZeroText.append(dataAlphabet.lookupObject(idCountPair.getID()) + " ");

        // Create a new instance named "test instance" with empty target and source fields.
        InstanceList testing = new InstanceList(instances.getPipe());
        testing.addThruPipe(new Instance(topicZeroText.toString(), null, "test instance", null));

        TopicInferencer inferencer = model.getInferencer();
        double[] testProbabilities = inferencer.getSampledDistribution(testing.get(0), 10, 1, 5);
        System.out.println("0\t" + testProbabilities[0]);