Table 3 Statistical significance analysis of the proposed method (SGA + RF) compared to baseline models.
Feature Subset Size | Proposed Method (SGA + RF) Accuracy (%) | Baseline Model (SVM) Accuracy (%) | Baseline Model (KNN) Accuracy (%) | Baseline Model (LR) Accuracy (%) | Paired t-test (p-value) | Statistical Significance (p < 0.05) |
|---|---|---|---|---|---|---|
8 | 93.67 | 88.45 | 87.12 | 85.79 | 0.032 | Significant |
12 | 97.71 | 91.29 | 90.65 | 89.98 | 0.018 | Significant |
16 | 98.54 | 92.84 | 91.45 | 90.23 | 0.022 | Significant |
22 | 99.01 | 93.17 | 92.41 | 91.56 | 0.015 | Significant |
26 | 96.32 | 91.75 | 90.98 | 89.12 | 0.027 | Significant |
30 | 95.19 | 90.33 | 89.77 | 88.66 | 0.031 | Significant |
34 | 96.69 | 91.08 | 90.22 | 88.98 | 0.029 | Significant |
