Fig. 5.

Should this be’Plots of classification accuracy and model complexity and of the combined measure EI. a The line plots of the classification model’s overall accuracy with the four wrapper algorithms on the 17 datasets, and the corresponding numbers of features selected by the different feature selection algorithms. b The line plots of the measurement EI. (Note that there is no relationship between the neighbouring datasets connected by the line which is simply included for convenience to identify data points for each algorithm. This is especially necessary where two algorithms have similar data)