An Optimized Neural Network Classification Method Based on Kernel Holistic Learning and Division

<div>Comparison of the classification performance and the number of generated training samples on the CC dataset. (a) Number of original training samples, number of generated training samples. (b) Number of original training samples, overall testing accuracy (%).</div>

Mathematical Problems in Engineering

fig10

Figure 10

Figure 10: An Optimized Neural Network Classification Method Based on Kernel Holistic Learning and Division 