Skip to main content
Fig. 2 | BMC Medical Informatics and Decision Making

Fig. 2

From: Evaluating semantic similarity methods for comparison of text-derived phenotype profiles

Fig. 2

A sample of an example output of the semantic similarity process. On the left is a semantic similarity matrix, in which every phenotype profile associated with a patient visit has been compared with every other one. The result is a matrix of similarity values. To evaluate the matrices, we then convert these into a ranked list of similarity values for each patient visit, which also includes whether or not the two patient visits being compared share a primary diagnosis. The latter structure is used to create our evaluation scores (e.g. AUC). In the experiment described in this article, this process is repeated once for every combination of semantic similarity measure being explored, since each will produce a separate similarity matrix

Back to article page