Change relevance labels and reorder documents to explore the effect the changes have on measures.
You can use the ir-measures package from Python or the command line to get the same results using:
ir_measures qrels run 'AP nDCG@10 P@10 P(rel=2)@10 R@10 R(rel=2)@10 Success@1 Success(rel=2)@1 Rprec RR ERR@10 RBP(rel=1) Judged@10'
ir_measures.calc_aggregate( [AP, nDCG@10, P@10, P(rel=2)@10, R@10, R(rel=2)@10, Success@1, Success(rel=2)@1, Rprec, RR, ERR@10, RBP(rel=1), Judged@10], ir_measures.read_trec_qrels('qrels'), ir_measures.read_trec_run('run'))
The following files are auto-generated and represent the scenario above.
This demo is powered by the ir-measures package, provided by the Terrier Team at the University of Glasgow.