Skip to main content
. Author manuscript; available in PMC: 2017 Jul 27.
Published in final edited form as: Proc ACM Int Conf Inf Knowl Manag. 2016 Oct;2016:297–306. doi: 10.1145/2983323.2983819

Table 3.

Performance results obtained for the system reported in this paper (Q/A-CDS) when using each type of medical sketch, method for answer inference, and relevance model as well as the iNDCG obtained by the State-of-the-Art (SotA) automatic and manual systems submitted to TREC.

iAP iNDCG R-Prec P@10
Q/A-CDS Z(t) .006 .010 .020 .062
EZ(t,l) .147 .434 .344 .722

Bethe Approximation .140 .432 .336 .701
Pair-wise .128 .382 .330 .610
Interpolated .147 .434 .344 .722

BM25 .042 .204 .163 .387
TF-IDF .041 .197 .169 .350
LMJM 0.040 .193 .151 .357
LMDir 0.043 .203 .170 .360
DFR .039 .197 .167 .333

SotA Task A Automatic .294
Task A Manual .311

Task B Automatic .382
Task B Manual .381