Table 6 Ablation experiments with 1% labeled samples.

From: Measured multi-source semi-supervised working condition recognition based on curvelet pooling and attention mechanism learning

Working condition recognition methods

Accuracy (%)

AMMFFE-CP-ResNet-50 (w/o FI)

87.88

AMMFFE-CP-ResNet-50 (w/ FI-AVE)

88.12

AMMFFE-CP-ResNet-50 (w/ FI-0.5FIX)

85.70

AMMFFE-CP-ResNet-50 (ours)

89.70

  1. Significant values are in bold
Back to article page