The Altair Community is migrating to a new platform to provide a better experience for you. The RapidMiner Community will merge with the Altair Community at the same time. In preparation for the migration, both communities are on read-only mode from July 15th - July 24th, 2024. Technical support via cases will continue to work as is. For any urgent requests from Students/Faculty members, please submit the form linked here.
Options

Term frequency vs TF-IDF

Member Posts: 9 Contributor II
edited November 2018 in Help
Hi friends,

I'm processing documents from files. Anyone know the diference between term frequency vector and tf-idf vector?

• Options
Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, Community Manager, RMResearcher, Member, University Professor Posts: 1,751 RM Founder
Hi,

term frequency is simply the normalized number of occurrences of a term divided by the number of terms of the document (from the souce code):

int numTerms = wordList.size();
double totalTermNumber = 0;
for (float value: frequencies)
totalTermNumber += value;

// Create the result structure
double[] wv = new double[numTerms];

// If document contains at least one term
if (totalTermNumber > 0) {
// Create the vector
double length = 0.0;
for (int i = 0; i < wv.length; i++) {
wv = frequencies / totalTermNumber;
length += wv * wv;
}

length = Math.sqrt(length);

// Normalize the vector
if (length > 0.0)
for (int i = 0; i < wv.length; i++)
wv = wv / length;
}

In case of TDIDF, the term frequency of each term is normalized with the inverse document frequency (from the source code):

// Obtain the total number of documents and the document frequencies
int numDocuments = wordList.getNumberOfDocuments();
int[] docFrequencies = wordList.getDocumentFrequencies();
double totalTermNumber = 0;
for (float value: frequencies)
totalTermNumber += value;

// Create the result structure
double[] wv = new double[docFrequencies.length];

// Create the vector

// If the document contains at least one term
if (totalTermNumber > 0) {
double length = 0.0;
for (int i = 0; i < wv.length; i++) {
// Note: docFrequencies is always > 0 as otherwise the word
// would not be in the word list, it is also always smaller as
// the total number of documents

double idf = Math.log(((double) numDocuments) / ((double) docFrequencies));
wv = (frequencies / totalTermNumber) * idf;
length += wv * wv;
}
length = Math.sqrt(length);

// Normalize the vector
if (length > 0.0)
for (int i = 0; i < wv.length; i++)
wv = wv / length;
}

You can find a discussion about the normalization somewhere in the forum if this is interesting for you.

Cheers,
Ingo

• Options
Member Posts: 9 Contributor II
Thank you so much Ingo!

On Text classification with one are more suitable? Thanks in advance!!
• Options
Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, Community Manager, RMResearcher, Member, University Professor Posts: 1,751 RM Founder
Hi,

as always in data mining: just test it. If one of both preprocessings works better for your data then just go for it. In general, however, TFIDF has outperformed term frequency alone on all practical settings I have experienced so far.

Cheers,
Ingo