Identifying Critical Drivers of Transportation Carbon Emissions

To ensure sufficient model training and reliable evaluation, the entire dataset was partitioned into a training set and a test set at an approximate ratio of 80% (training set) to 20% (test set). This division balances the needs for model learning and validation: the training set, comprising 491 samples, provides the Random Forest with sufficient data to construct a robust ensemble of decision trees; the test set, containing 123 samples, remains statistically significant and effectively evaluates the model's generalization performance on unseen data.
Regarding model hyperparameter settings, to enhance the reproducibility of re-sults, the random seed was fixed at 131 for all analyses. Furthermore, a systematic optimization of model hyperparameters was conducted using a combination of grid search and cross-validation, selecting the parameter combination with the lowest cross-validation error as the optimal configuration. The core parameters determined through optimization are as follows: max_depth = 15, max_features = 9, and n_estimators = 300.