a simple implementation of TF-IDF algorithm in Java.

TFIDFCalculator.java

package com.guendouz.textclustering.preprocessing;

import java.util.Arrays;
import java.util.List;

/**
 * @author Mohamed Guendouz
 */
public class TFIDFCalculator {
    /**
     * @param doc  list of strings
     * @param term String represents a term
     * @return term frequency of term in document
     */
    public double tf(List<String> doc, String term) {
        double result = 0;
        for (String word : doc) {
            if (term.equalsIgnoreCase(word))
                result++;
        }
        return result / doc.size();
    }

    /**
     * @param docs list of list of strings represents the dataset
     * @param term String represents a term
     * @return the inverse term frequency of term in documents
     */
    public double idf(List<List<String>> docs, String term) {
        double n = 0;
        for (List<String> doc : docs) {
            for (String word : doc) {
                if (term.equalsIgnoreCase(word)) {
                    n++;
                    break;
                }
            }
        }
        return Math.log(docs.size() / n);
    }

    /**
     * @param doc  a text document
     * @param docs all documents
     * @param term term
     * @return the TF-IDF of term
     */
    public double tfIdf(List<String> doc, List<List<String>> docs, String term) {
        return tf(doc, term) * idf(docs, term);

    }

    public static void main(String[] args) {

        List<String> doc1 = Arrays.asList("Lorem", "ipsum", "dolor", "ipsum", "sit", "ipsum");
        List<String> doc2 = Arrays.asList("Vituperata", "incorrupte", "at", "ipsum", "pro", "quo");
        List<String> doc3 = Arrays.asList("Has", "persius", "disputationi", "id", "simul");
        List<List<String>> documents = Arrays.asList(doc1, doc2, doc3);

        TFIDFCalculator calculator = new TFIDFCalculator();
        double tfidf = calculator.tfIdf(doc1, documents, "ipsum");
        System.out.println("TF-IDF (ipsum) = " + tfidf);


    }


}

Thanks mate, couldn't figure out how to compute this efficiently enough !

Thanks. Just starting on this toping and this is great help.

C:\Users\M\Desktop>javac TFIDFCalculator.java
C:\Users\M\Desktop>java TFIDFCalculator
Error: Could not find or load main class TFIDFCalculator
Sir, I am getting this error, can you help me to sort out this error..

K75: It sounds like you have %CLASSPATH% defined, given no other information. Try:

javac TFIDFCalculator.java
java -cp . TFIDFCalculator

in idf method you should add the case where "term" doesn't exist because you can't /0

return Math.log(docs.size() / n);
cast it..
return Math.log((double)docs.size() /*(double) n);

Everything allrigth

Thanks.

This is a great demo, thanks for putting this up

Thank you, just getting into text mining and this is very helpful.

Hi,

I'm currently looking into TF-IDF for the first time.
One detail question:

In line 39 you are using Math.log, which returns the natural logarithm (base e) (https://docs.oracle.com/javase/7/docs/api/java/lang/Math.html).
The wikipedia article (https://en.wikipedia.org/wiki/Tf%E2%80%93idf) states that base 10 logarithm should be used, so shouldn't this be changed to:
Math.log10(docs.size() / n)

Kind regards,
Michael

I think you will get exception in line return Math.log(docs.size() / n); in case n is 0.