Computational Precision Health

by on

Development of a machine learning-derived model to predict unplanned ICU admissions after major non-cardiac surgery

Model summary of the most influential variables. Model summary of the most influential variables. Each dot represents an individual case-variable pair. The x-axis shows the SHAP value (impact on log-odds of UIA): dots to the right increase predicted risk, dots to the left decrease it. Red dots indicate higher numeric value, and blue dots indicate lower numeric value. Dots (regardless of color) on the right side of the y-axis indicate a higher impact value on the model to predict a positive UIA, and vice versa. Dots with the same impact value are plotted on top of each other to create a vertical distribution