Fig. 10 | Scientific Reports

Fig. 10

From: Convolutional neural network approach for automated well zonation in the Lower Bahariya member north Western Desert Egypt

Fig. 10

Permutation feature importance derived from the convolutional neural network (CNN) classification model. Each bar represents the mean decrease in model accuracy when the corresponding feature is randomly permuted, averaged over 100 repetitions. Features causing higher accuracy drops are considered more influential on the model’s predictive performance.

Back to article page