The Main Objective of this project is to create a Learning to Rank system which will eventually be integrated in Xapian. This project can be broken down into the following components.
- Optimizing the existing code.
- Integrating the remaining rankers (ListMLE and AdaRank) and scorers (ERR scorer) from vhasu's 2014 work.
- An automated test suite for the existing Letor code, covering both the higher-level API and the lower-level pieces.
- Revising existing Documentation and writing practical code examples.
- Testing for performance and scale with INEX2009 or similar data-set by using Letor.