Table 9 Comparison of the Proposed approach with RF and KNN classifiers outperforms the state-of-the-art MS detection methods conducted on the eHealth lab dataset, achieving top accuracy and sensitivity. Specifically, the combination with KNN reaches the highest marks, showcasing the effectiveness of this method in MS detection.

HWT + PCA + LR⁶¹

89.72

-

WE + FNN + AGA⁴⁴

91.95

91.91

HMI + FNN + PSO⁴⁶

91.70

91.67

GLCM + FNN⁴⁵

92.75

92.75

MAMFM⁶²

93.83

94.08

GLCM + ensemble + LogitBoost⁴¹

94.91

95.79

GLCM + GLRLM⁶³

95.14

95.27

6l-CNN⁶⁴

95.82

95.98

DWT + PCA + LS-SVM⁶⁵

96.21

95.86

WE + HBP⁶⁶

96.72

96.15

DWT + PPCA + RF⁶⁷

96.40

96.01

FRFE + MLP + ST-Jaya¹⁶

97.39

97.40

BWT + RKPCA + LR³⁸

97.76

97.12

MBD + SHLNN + TSR-BBO⁶⁸

97.80

97.78

SWE + KNN³⁷

97.94

96.15

BWF + FAGA⁶⁹

97.89

98.00

GLCM + LBP + Hybrid SHOSCA + RF (ours)

97.77

97.73

GLCM + LBP + Hybrid SHOSCA + KNN (ours)

97.97

98.89