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

Fig. 2

From: Explainable AI enables clinical trial patient selection to retrospectively improve treatment effects in schizophrenia

Fig. 2

Individual PAI prediction results for actual randomized arms. The plot shows the measured total PANSS score at week 6 vs. the averaged predicted total PANSS score at week 6 from the Elastic Net regression model (each dot is an individual patient, n = 197). Patients are colored by their actual randomized arm (paliperidone treatment in blue, placebo in orange), and as expected the week 6 scores are generally higher for patients receiving placebo. The dashed line is y = x. Variance explained by the model is 58%. Note that these are not the counterfactual predictions

Back to article page