• In information retrieval, a weighting procedure so that some words in a query or document get emphasized more than others. A document is ranked higher using TF-IDF when it has more occurrences of the query term (TF or term frequency) and ranks lower when the word occurs in more documents (IDF or inverse document frequency). There are different rules for deciding how to combine TF with IDF, on common rule is to rank the documents based on the ratio of TF to log(IDF).


  1. Herb Roitblat, Search 2020: The Glossary.
Print Friendly, PDF & Email