Artificial Neural Network Individualised Prediction of Time to Colorectal Cancer Surgery

<div>(a) ROC for the eight-week analysis. AUC for prediction of time to surgery was 0.793. (b) Gain chart shows ANN utility of predicting times longer than eight weeks. (c) Lift chart. Using 10% of the cohort, the ANN was 3.8 times more likely to correctly predict time to surgery over random selection. DtL: diagnosis to laparoscopy.</div>

Gastroenterology Research and Practice

fig1

Figure 1

Figure 1: Artificial Neural Network Individualised Prediction of Time to Colorectal Cancer Surgery 