Table 4 Comparative analysis with related works.
Study | Method | Focus | Interpretability | Key findings |
|---|---|---|---|---|
This study | RF + SHAP + PDP | Tomato fruit expansion | High | Soil temp 21.8 °C, light > 20 Klux, soil EC 0.6–0.8 dS/m |
[35] Yu et al., 2024 | ANN + SSA optimization | Fertilizer application rate prediction for tomato | Medium | ANN outperformed traditional models in NPK prediction, but limited interpretability for physiological mechanisms |
[36] Zhang et al., 2022 | XGBoost regression | Greenhouse tomato evapotranspiration prediction | Medium | XGBR-ET achieved superior accuracy, supporting irrigation scheduling, though without feature-level interpretation |
[37] Wang et al., 2025 | XGBoost | Irrigation prediction for cherry tomato | Low | Enhanced precision irrigation through XGBoost, though remained a black-box approach |
[38] Mancer et al., 2024 | ML (RF, XGBoost) | Tomato yield prediction in greenhouse | Medium | Improved yield prediction accuracy, but lacked explainability of environmental–yield mechanisms |
