Part of research on implementing quantum support vector machines on multi-dimensional graph based data to preserve its topogical strucutre for more meainingful classification. Mainly trained on TUDataset's PROTEINS and MUTAG datasets for the classification of biological molecules.
Scored by 10-Fold Cross-Validation
PROTEINS - 74.48% ± 3.75
MUTAG - 83.48% ± 6.61
- Preserves topological structure of graph data
- Consistent sub 3:00 training time
- Relatively simple feature map
- Highly interpretable model
The included ipynb file only includes the code regarding the final model.