Skip to main content
Annals of Intensive Care
Fig. 6 | Annals of Intensive Care

Fig. 6

From: Automated characterization of patient–ventilator interaction using surface electromyography

Fig. 6

Patient–ventilator interaction classification metrics. For the triangle algorithm (

figure i
) and the adaptive thresholding algorithm (
figure j
), the six classes and 32 recordings, the sensitivity, specificity and positive predictive value are evaluated. This plot visualizes the distribution of the three metrics over the datasets. Black lines indicate the median value and white lines mark the interquartile range. In addition, the results of the different classes are summarized on the right side, where Total denotes the average over all classes and in Total (weighted) considers the class frequency. In this way, the latter represents a reality-oriented outcome

Back to article page