Classification Modeling Method for Near-Infrared Spectroscopy of Tobacco Based on Multimodal Convolution Neural Networks

Table 3

Classification performance of 1-D CNN models with different depths.

Model	Depth	Training set classification accuracy (%)	Test set classification accuracy (%)
1D_CNN_5	5	82.75	74.71
1D_CNN_7	7	88.1	81.28
1D_CNN_9	9	92.46	87.05
1D_CNN_11	11	95.56	90.81
1D_CNN_13	13	96.72	88.13