Artificial image objects for classification of breast cancer biomarkers with transcriptome sequencing data and convolutional neural network algorithms

Table 2 Cross-validation and sample testing results for Ki67

	Accuracy	AUC	Precision	Recall	F1-score
Cross-validation
Ki67-			0.815 ± 0.036	0.834 ± 0.022	0.824 ± 0.026
Ki67 +	0.821 ± 0.023	0.891 ± 0.021	0.831 ± 0.016	0.811 ± 0.036	0.820 ± 0.024
Weighted average	0.821 ± 0.023	0.891 ± 0.021	0.822 ± 0.023	0.822 ± 0.024	0.822 ± 0.024
Sample testing on GSE81538
Ki67-			0.886 ± 0.010	0.796 ± 0.048	0.837 ± 0.021
Ki67 +	0.826 ± 0.037	0.883 ± 0.016	0.763 ± 0.036	0.864 ± 0.021	0.809 ± 0.016
Weighted average	0.826 ± 0.037	0.883 ± 0.016	0.833 ± 0.021	0.825 ± 0.036	0.825 ± 0.019

ISSN: 1465-542X