Skip to content

README.md

Latest commit

 

History

History
26 lines (22 loc) · 1.23 KB

README.md

File metadata and controls

26 lines (22 loc) · 1.23 KB

SimpleIRSystem

A simple IR system to index and query a collection.

Collaborators

  • Charles Du
  • Michael Kyeyune

Requirements

  • Python 3.4.3+

Utilising this system

To utilise this system you need to generate a collection from a testbed, index the collection then compare MAP/Avg NDCG results for the modified and unmodified engine.

  • generate collection from testbed
    • python collect.py testbedx - x being the number of the testbed. This will generate a file testbedx_collection
  • index testbed collection
    • python index.py testbedx_collection
  • analyse modified and unmodified engine performance
    • python analyse.py testbedx
  • finding optimal indicative terms and top k documents to utilise for blind relevance feedback (BRF) for a single testbed
    • python optimise.py -s testbedx 200 - 200 being the number of documents to consider in MAP/Avg NDCG calculations
  • finding optimal indicative terms and top k documents to utilise for BRF for all testbeds
    • python optimise.py -a 200

Important Notices

  • Documents that are not in UTF-8 format are ignored when generating the collection for a testbed
  • All testbeds must be indexed before attempting to optimise across all testbeds