Fig 1.
The workflow of the submitted proposal.
Table 1.
Summary of monthly air quality and weather parameters in Paris (2023).
Fig 2.
Heat map correlation of different Air pollutant parameters.
Table 2.
Summarization of outlier and missing values in the dataset.
Table 3.
Hyperparameter optimization setup and final values.
Table 4.
Performance comparison of meta-learners in stacking (validation set).
Table 5.
Performance metrics for LSTM deep learning benchmark model.
Fig 3.
Actual vs. predicted values for LSTM Model on PM2.5 and CO.
Table 6.
Performance metrics for air pollution forecasting models.
Table 7.
Baseline model comparisons.
Fig 4.
Model predictions vs. actual pollutant values over a sample time series (n = 50).
Fig 5.
Actual vs. predicted scatter plots.
Fig 6.
Residual plots for NO prediction: Stacked Ensemble residuals cluster tightly around zero.