Table 21 Performance comparison of patternfusion with and without contrastive pretraining.
From: PatternFusion: a hybrid model for pattern recognition in time-series data using ensemble learning
Model | Dataset | Accuracy (%) | F1-score (%) | AUC (%) |
|---|---|---|---|---|
PatternFusion (baseline) | PUT | 96.4 | 94.9 | 98.1 |
PolyU | 95.8 | 93.7 | 97.6 | |
BioCop | 94.7 | 92.5 | 96.3 | |
PatternFusion + contrastive pretraining | PUT | 97.2 | 95.6 | 98.7 |
PolyU | 96.6 | 94.4 | 98.3 | |
BioCop | 95.4 | 93.1 | 97.1 |
