TF.IDF
We sometimes use log() for term frequency. Does this:
The more a term is observed in the corpus, the
Some search engines will remove terms that occur in more than 50% of documents. Is this because: