Quiz
Which of the following is true?
- Jelinek-Mercer language modeling involves taking the language model for a document and then pretending
to add `mu` documents to this model according to an underlying language model for the corpus.
- The derivation of the LMD relevance measure makes use of random walks through documents.
- Naively partitioning the index by term is more likely to create node hot-spots than partitioning the index by document.