Relevance Models are a classic probabilistic version of blind relevance feedback (BRF) based approach for language modeling IR.
- Estimates a query language model
P ( t | q )based on top results - Assumes the top
kranked results by query likelihood (QL) as relevant
Re-rank documents by P(Q*|D,R) --> A4 (Product of P(t | D, R)), t is expanded query terms (query terms + expanded
top n terms from documents).
Lavrenko, V., & Croft, W. B. (2001). Relevance based language models. In Proceedings of the 24th annual international ACM SIGIR conference on Research and development in information retrieval. SIGIR01: 24th ACM/SIGIR International Conference on Research and Development in Information Retrieval. ACM. https://doi.org/10.1145/383952.383972