Answers are grounded in this paper and cite their sources.
CO 2 emissions from biomass hydrothermal liquefaction (HTL) are a critical environmental concern, particularly given the need to optimize feedstock selection and process parameters for climate‐friendly bioenergy production. This study aimed to develop robust machine learning (ML) models for accurate prediction of CO 2 emissions based on comprehensive input data encompassing feedstock elemental and biochemical composition as well as HTL operating conditions. A curated dataset of 613 high‑quality literature records was compiled and preprocessed, with 90% used for training and 10% for independent validation. Eight ML algorithms, Decision Tree, AdaBoost, Random Forest, KNN, Ensemble Learning, CNN, SVR, and MLP‑ANN, were evaluated using 5‐fold cross‐validation and performance metrics (R 2 , MSE, AARE%). Hyperparameter tuning was conducted individually for each algorithm to maximize generalization. Results identified Support Vector Regression (SVR) and Random Forest as the most effective models, exhibiting high and consistent R 2 values (∼0.93) across all data splits with low error metrics, while Decision Tree suffered from overfitting despite perfect training accuracy. SHAP analysis revealed reaction temperature as the dominant predictor, exerting a strong positive effect on emissions, followed by oxygen content with a negative influence; other variables exhibited mixed or weaker impacts. These findings highlight that precise control of key HTL parameters—especially thermal intensity—can substantially influence CO 2 outputs. The integration of interpretability analytics ensures that predictive models not only offer accurate forecasts but also guide process optimization to mitigate emissions.
No takes yet. Share an insight, caveat, or question.
Yadav et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: