Table 4 Comparison of different approaches in terms of key factors
Criterion | Zero-shot BERT | Generative LLM | Rule-based | MISTIC |
|---|---|---|---|---|
Training data | No | No | Predefined rules | Few |
Explainability | No | No | Yes | Yes |
Generalization | Yes | Yes | No | Yes |
Computational demand | Moderate | High | Low | Moderate |
Automation | Full | Full | Partial | Full |
Manual effort | Low | Low | High | Low |