Give an example of a conditional and a joint probability using the occurrence of
words in documents as the events.
7.5. Implement a BM25 module for Galago. Show that it works and document
it.
7.6. Show the effect of changing parameter values in your BM25 implementation.
7.7. What is the “bucket” analogy for a bigram language model? Give examples.
7.8. Using the Galago implementation of query likelihood, study the impact of
short queries and long queries on effectiveness. Do the parameter settings make a
difference?
7.9. Implement the relevance model approach to pseudo-relevance feedback in
Galago. Show it works by generating some expansion terms for queries and document
it.
7.10. Show that the belwand operator computes the query likelihood score with
simple terms. What does the belwsum operator compute?
7.11. Implement a #not operator for the inference network query language in
Galago. Show some examples of how it works.
7.12. Do a detailed design for numeric operators for the inference network query
language in Galago.
7.13. Write an interface program that will take a user’s query as text and transform
it into an inference network query. Make sure you use proximity operators.
Compare the performance of the simple queries and the transformed queries