Concrete Spalling Severity Classification Using Image Texture Analysis and a Novel Jellyfish Search Optimized Machine Learning Approach

Table 2

Experimental result comparison.

Phase	Metrics	JSO-SVC		RFC		CNN-Adam		CNN-RMSprop
		Mean	Std	Mean	Std	Mean	Std	Mean	Std
Training	CAR (%)	95.926	0.574	98.500	0.441	90.278	3.341	89.056	5.207
	Precision	0.950	0.008	0.978	0.005	0.932	0.052	0.915	0.062
	Recall	0.969	0.007	0.993	0.006	0.875	0.079	0.870	0.113
	NPV	0.969	0.007	0.993	0.006	0.887	0.060	0.886	0.076
	F1 score	0.960	0.006	0.985	0.004	0.899	0.039	0.885	0.068
	AUC	0.983	0.002	0.998	0.001	0.966	0.025	0.952	0.039
Testing	CAR (%)	93.333	3.801	87.500	4.941	81.500	5.669	79.167	8.298
	Precision	0.932	0.048	0.871	0.070	0.877	0.090	0.809	0.104
	Recall	0.936	0.062	0.890	0.069	0.750	0.118	0.777	0.136
	NPV	0.935	0.057	0.892	0.062	0.788	0.076	0.794	0.098
	F1 score	0.933	0.042	0.877	0.047	0.799	0.067	0.785	0.096
	AUC	0.969	0.020	0.944	0.046	0.896	0.060	0.855	0.080