Empathy Detection on HRI (5-fold 10-repeat stratified cross-val)
71.5AccuracyTFMPathyFT(TabPFN)
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| TFMPathyFT(TabPFN)Evaluation Protocol=Fine-Tuning, Backbone/Foundation Model=TabPFN, Feature Type=Summary features2025.04 | 71.5 | 70 | 77.3 | 65.9 | 69.1 | |
| XGBoostImplementation Source=Reported in literature, Feature Type=Summary features2025.04 | 69 | 72 | 69 | 69 | 69 | |
| TFMPathyICL(TabICL)Evaluation Protocol=In-Context Learning, Backbone/Foundation Model=TabICL, Feature Type=Summary features2025.04 | 63.4 | 66.5 | 63.7 | 66 | 63.9 | |
| RFFeature Type=Summary features2025.04 | 62.5 | 65.1 | 63.1 | 63.9 | 62.8 | |
| TFMPathyICL(TabPFN)Evaluation Protocol=In-Context Learning, Backbone/Foundation Model=TabPFN, Feature Type=Summary features2025.04 | 62.2 | 65.3 | 62.1 | 65.1 | 62.7 | |
| XGBoostHyperparameter Configuration=Optuna optimisation, Feature Type=Summary features2025.04 | 61.2 | 65.1 | 61.1 | 63.8 | 61.5 | |
| SVMFeature Type=Summary features2025.04 | 60.5 | 64.1 | 61.9 | 58.1 | 59.2 | |
| XGBoostImplementation Source=Our re-implementation, Feature Type=Summary features2025.04 | 59.8 | 62.3 | 60.4 | 58.9 | 58.8 |