Table 7.
List of XAI studies performing explanation consistency assessment
References | # Cit. | Application | Input Data | AI model(s) | XAI method(s) | Dataset(s) |
---|---|---|---|---|---|---|
Thimoteo et al. (2022) | 2 | COVID-19 diagnosis | EHR | SVM, RF | SHAP | COVID-19 Data Sharing/BR |
Alves et al. (2021) | 36 | COVID-19 diagnosis | EHR | RF | DTX, Criteria graph, LIME, SHAP | COVID-19 dataset |
Okay et al. (2021) | 1 | Diabetes diagnosis | EHR | RF, GBDT | SHAP, LIME | Sylhet Diabetes dataset |
Oba et al. (2021) | 1 | Diabetes diagnosis | EHR | TabNet, XGBoost, LightGBM, CatBoost | SHAP (all), attention (TabNet) | Retrospective study |
Elshawi et al. (2019) | 132 | Hypertension prediction | EHR | RF | feature permutation, PDP, ICE, global surrogate models, LIME, SHAP | Pilot study |
Seedat et al. (2020) | 1 | Voice pathology assessment | Audio features | ExtraTrees | SHAP,Morris sensitivity analysis | Pilot study |
Kapcia et al. (2021) | 0 | Lung cancer life expectancy prediction | EHR | RF | LIME, SHAP | Simulacrum dataset |
Duell et al. (2021) | 9 | Lung cancer mortality prediction | EHR | XGBoost | LIME, SHAP, Anchors | Simulacrum dataset |
Moncada-Torres et al. (2021) | 43 | Breast cancer survival prediction | EHR | XGBoost | SHAP | Retrospective study |
Ang et al. (2021) | 0 | ICU mortality risk prediction | EHR | RF, MLP | SHAP | MIMIC-III |
Song et al. (2020) | 33 | AKI prediction | EHR | GBDT | SHAP | Retrospective study |
Duckworth et al. (2021) | 5 | Hospital readmission prediction | EHR | XGBoost | SHAP | Retrospective study |
Tahmassebi et al. (2020) | 3 | Eye state detection | EEG | XGBoost, DNN | SHAP | Pilot study |
Antoniadi et al. (2021) | 7 | QoL assessment in ALS caregiving | EHR | XGBoost | SHAP | Retrospective study |
Ward et al. (2021) | 5 | Pharmaco-vigilance monitoring | EHR | RF, XGBoost, ExtraTrees | MDA, MDI, LIME, SHAP | Retrospective study |