Fig. 3From: Clinical language search algorithm from free-text: facilitating appropriate imagingAccuracy on simulated indications dataset. This figure is a cumulative graph of the percentage of indications with ground truths ranked as the top search result to those within the top 60 search results. It shows a significant difference between algorithm performances on simple and complex simulated indications (p < 1e−10, two sample Kolmogorov Smirnov test)Back to article page