Blended Features Classification of Leaf-Based Cucumber Disease Using Image Processing Techniques
Table 7
Performance evaluation of experiment 3.
Classifier
Performance evaluation
Accuracy (%)
Specificity (%)
Sensitivity (%)
Precision (%)
FNR (%)
FPR
Subspace KNN
94.2
82.5
82.5
83.66
5.8
0.011
Bagged trees
93.0
95.2
94.8
79.16
7.0
0.015
Fine KNN
93.9
83.1
82.1
83.66
6.1
0.013
Cubic SVM
91.8
93.6
91.0
94.00
8.2
0.018
Fine Gaussian SVM
71.2
62.2
62.2
81.8
28.2
0.068
Fine tree
82.1
77.6
77.6
83.2
17.9
0.08
Weighted KNN
93.4
95.6
90.7
96.9
6.6
0.016
Bold shows the highest values in terms of accuracy. The subspace classifier has the highest accuracy of 94.2%, specificity of 82.5%, sensitivity of 82.5%, precision of 83.66%, FNR rate of 5.8, and FPR is 0.011.
