Fig. 1

Epoch-level classification metrics for the classification of epochs as sitting (yes/no) by CHAP-child (N = 202 participant-seasons from 84 participants). Each metric was calculated for each participant-season and values at the top of the chart reflect the mean (SD) across participant-seasons. CHAP = Convolutional Neural Network Hip Accelerometer Posture classification method; SD = standard deviation; Balanced Accuracy = average of sensitivity and specificity; Sensitivity = correctly classified as sitting (i.e., true positives) / actual sitting; Positive Predictive Value = correctly classified as sitting (i.e., true positives) / classified sitting; Specificity = correctly classified as non-sitting (i.e., true negatives) / actual non-sitting; Negative Predictive Value = correctly classified as non-sitting (i.e., true negatives) / classified non-sitting