Fig. 10From: Development and validation of ‘Patient Optimizer’ (POP) algorithms for predicting surgical risk with machine learningSHAP visualisation for a specific patient’s length-of-stay. This is a True Positive (TP) prediction of a high length-of-stay, with a probability of 0.72. The length of the bar indicates the influence of that feature on the prediction. The colour indicates whether the influence is positive (red) or negative (blue). The grey value to the left of the feature name is the value of that feature for this patientBack to article page