Table 3 Comparison of crystal system classification accuracy between PE-AG-GMoE and RF Models
From: Accelerating electron diffraction analysis using graph neural networks and attention mechanisms
Crystal System | PE-AG-GMoE | RF | PE-AG-GMoE (Agg.) | RF (Agg.) |
|---|---|---|---|---|
Cubic | 97.8 | 92 | 98.9 | 94 |
Hexagonal | 93.0 | 78 | 97.1 | 93 |
Trigonal | 84.6 | 86 | 91.8 | 98 |
Tetragonal | 93.3 | 70 | 98.1 | 85 |
Orthorhombic | 92.3 | 57 | 99.0 | 71 |
Monoclinic | 84.9 | 62 | 96.2 | 72 |
Triclinic | 64.5 | – | 83.8 | – |
Accuracy (avg.) | 89.2 | 74.2 | 96.5 | 85.5 |
