Fig. 3From: A clustering approach for detecting implausible observation values in electronic health records dataChanges in specificity of the clustering approach by αBack to article page