Fig. 6From: Deep learning for DNase I hypersensitive sites identificationAnalysis of Arabidopsis, rice, and human data shows that the number of DHSs decreases with increasing length. However, in the benchmark dataset, because of the tiny size, it cannot display the trend. It is embarrassing to enhance the prediction capacity of the model, so it is easy to fall into overfittingBack to article page