Table 1.
Summary of existing literature.
Table 2.
Summary statistics.
Fig 1.
Pair plot of all attributes with respect to number of uniform barriers.
Fig 2.
Pair plot of all attributes with respect to number of Gaussian barriers.
Fig 3.
Correlation heatmap of all attributes.
Fig 4.
(a) Histogram of Number of Gaussian Barriers and (b) Histogram of Number of Uniform Barriers.
Table 3.
Algorithm for hyperparameter tuning with ACO.
Fig 5.
Ranking of features according to feature importance.
Table 4.
Algorithm for feature importance analysis.
Table 5.
Algorithm for regularization techniques application.
Table 6.
Algorithm for feature sensitivity analysis.
Table 7.
Initial regression model results.
Fig 6.
(a) Best Solution for ACO–SVR1 Model and (b) Best Solution for ACO–SVR2 Model.
Table 8.
Results for ACO–SVR1 and ACO–SVR2 Model.
Table 9.
Percentage Improvement in ACO–optimized models compared to initial models.
Table 10.
Algorithm of ACO–SVR hyperparameter optimization.
Fig 7.
Scatter plot of actual vs. predicted values for SVR1 model.
Fig 8.
Scatter plot of residual vs actual values for SVR1 model.
Fig 9.
Scatter plot of actual vs. predicted values for SVR2 model.
Fig 10.
Scatter plot of residual vs actual values for SVR2 model.
Fig 11.
(a) Best Distance Over Iterations using ACO–SVR1 Model and (b) Best Distance Over Iterations using ACO–SVR2 Model.
Fig 12.
(a) Best Solution and All Nodes for ACO–SVR1 Model and (b) Best Solution and All Nodes for ACO–SVR2 Model.
Fig 13.
Scatter plot of actual vs. predicted values for ACO–SVR1 model.
Fig 14.
Scatter plot of residual vs actual values for ACO–SVR1 model.
Fig 15.
Scatter plot of actual vs. predicted values for ACO–SVR2 model.
Fig 16.
Scatter plot of residual vs actual values for ACO–SVR2 model.
Fig 17.
Scatter plot to compare SVR–1 with ACO–SVR1 model.
Fig 18.
Scatter plot to compare SVR with ACO–SVR1 model.
Table 11.
Summary of the results of hyperparameter tuning using ACO–SVR1 and ACO–SVR2 models.
Fig 19.
(a) Scatter Plot of Actual vs Predicted Values of Number of Barriers for Model 1 and (b) for Model 2.
Fig 20.
(a) Plot of Residuals for Number of Barriers for Model 1 and (b) for Model 2.
Fig 21.
Bar plot comparing the effect of L1 and L2 regularization on various metrics for model 1.
Fig 22.
Plot comparing the effect of L1 and L2 regularization on various metrics for model 2.
Fig 23.
(a) Scatter Plot to Compare the Initial vs Final Predictions for Model 1 (b) and for Model 2.
Fig 24.
Bar plot illustrating the percentage improvement in performance metrics of final models.