Table 2 Corrected data results per model.

From: System transferability of Raman-based oesophageal tissue classification using modern machine learning to support multi-centre clinical diagnostics

 

5 × 3 CV

one centre

Centre 1

Centre 2

Centre 3

5 × 3 CV

All centres

PCA-LDA

Centre 1

62.7% +/− 4.3

70.4%

70.5%

73.7%

62.2% +/− 11.2

Centre 2

64.4% +/− 8.3

67.2%

75.4%

68.9%

Centre 3

63.4% +/− 9.7

68.9%

67.2%

73.8%

SVM

Centre 1

64.1% +/− 13.7

80.3%

63.9%

75.4%

66.6% +/− 8.8

Centre 2

64.0% +/− 12.9

62.3%

80.3%

68.9%

Centre 3

65.1% +/− 12.1

62.2%

65.6%

78.7%

CNN

Centre 1

63.0% +/− 11.1

100%

54.1%

78.7%

68.2% +/− 9.6

Centre 2

61.4% +/− 11.3

62.3%

100%

65.6%

Centre 3

61.1% +/− 11.8

68.9%

59.0%

93.4%

  1. Top line per centre represents accuracy (%) +/− 1 SD, bottom line represents log loss +/− 1 SD. Italicised entries indicate the training results and so are not indicative of performance.