Figure 4.

(A) The impact of the window size on the model’s performance. (B) The performance comparison of twenty handcrafted features and adaptive learning embedding features on four traditional machine learning classifiers