Fig. 5.

Stacked Ensemble model to distinguish common and uncommon EGFR mutations. A–C Discrimination of Stacked Ensemble model for identification of common and uncommon EGFR mutations in the training cohort, the testing cohort and total patients, respectively. D The importance of differential clinical parameters in the Stacked Ensemble model. The AutoML randomly generated twenty algorithms based on deep learning model and five machine learning model, and the twenty models were interpreted using feature importance plots through the matplotlib package built in the software