Cross-Topic Learning for Work Prioritization in Systematic Review Creation and Update

Table 5

Mean AUC Performance Comparison of the Three Systems Studied Averaged Across All Topics. Statistical Significance Shown for Pairs of Systems at the Different Amounts of Training Data Represented by Differing Levels of Reverse Cross-Validation. p Values Computed Using the Nonparametric Paired Wilcoxon Test, Comparing Pairs of System Performance on Each of the 24 SR Topics

	Average Mean AUC Across All Topics		Topic Pairwise Wilcoxon Test	Average Mean AUC Across All Topics		Topic Pairwise Wilcoxon Test
XVAL	Hybrid	Baseline	p-value	Hybrid	Non-Topic	p-value
2	0.900	0.896	0.020	0.900	0.795	0.000
4	0.879	0.870	0.002	0.879	0.795	0.000
8	0.860	0.841	0.000	0.860	0.795	0.000
16	0.841	0.807	0.000	0.841	0.795	0.000
32	0.826	0.773	0.000	0.826	0.795	0.005
64	0.811	0.727	0.000	0.811	0.795	0.160
128	0.796	0.675	0.000	0.796	0.795	0.574

AUC = area under the curve; SR = systematic review.

A p-value of 0.0000 Indicates p < 0.00005.

This Article

Access policy for JAMIA

All content published in JAMIA is deposited with PubMed Central by the publisher with a 12 month embargo. Authors/funders may pay an Unlocked fee of $2,000 to make the article free on the JAMIA website and PMC immediately on publication.

All content older than 12 months is freely available on this website.

AMIA members can log in with their JAMIA user name (email address) and password or via the AMIA website.