Skip to Main Content
This program has achieved 91.12% accuracy on the test set, which is the best performance out of the 300 independent runs. Table 4 shows the confusion matrix for this program on the test set. The program has perfectly classified the instances of 10 classes, scored over 80% on 16 classes, and performed poorly on two classes (rug1 and wall) only, as depicted in Table 4. The program did struggle with generating a good feature vector for the rug1 class as it achieved only 46% accuracy. The majority of those misclassified instances from rug1 are confused with grass1. A closer inspection of a sample from each of these two classes, that is, rug1 and grass1, shows that both instances share similar characteristics such as the structure of the grass leaves as presented in Figure 26.
Figure 26:

Samples taken from the KyNoRo dataset.

Figure 26:

Samples taken from the KyNoRo dataset.

Table 4:

Confusion matrix for the best evolved program on the KyNoRo dataset, where the first row lists the predicted labels based on the generated feature vectors by this program and using 1-NN, and the actual class labels are listed in the first column.

blanket1blanket2canvas1ceiling1ceiling2cushion1floor1floor2grass1lentils1linseeds1oatmeal1pearlsugar1rice1rice2rug1sand1scarf1scarf2screen1seat1seat2sesameseeds1stone1stone2stone3stoneslab1wall1Accuracy
blanket1 69 11 0.86 
blanket2 74 0.93 
canvas1 80 1.00 
ceiling1 66 14 0.83 
ceiling2 80 1.00 
cushion1 80 1.00 
floor1 80 1.00 
floor2 80 1.00 
grass1 11 64 0.80 
lentils1 80 1.00 
linseeds1 80 1.00 
oatmeal1 74 0.93 
pearlsugar1 71 0.89 
rice1 79 0.99 
rice2 80 1.00 
rug1 23 37 0.46 
sand1 74 0.93 
scarf1 80 1.00 
scarf2 80 1.00 
screen1 74 0.93 
seat1 79 0.99 
seat2 72 0.90 
sesameseeds1 71 0.89 
stone1 79 0.99 
stone2 69 0.86 
stone3 65 0.81 
stoneslab1 74 0.93 
wall1 17 11 50 0.63 
Average  0.91 
blanket1blanket2canvas1ceiling1ceiling2cushion1floor1floor2grass1lentils1linseeds1oatmeal1pearlsugar1rice1rice2rug1sand1scarf1scarf2screen1seat1seat2sesameseeds1stone1stone2stone3stoneslab1wall1Accuracy
blanket1 69 11 0.86 
blanket2 74 0.93 
canvas1 80 1.00 
ceiling1 66 14 0.83 
ceiling2 80 1.00 
cushion1 80 1.00 
floor1 80 1.00 
floor2 80 1.00 
grass1 11 64 0.80 
lentils1 80 1.00 
linseeds1 80 1.00 
oatmeal1 74 0.93 
pearlsugar1 71 0.89 
rice1 79 0.99 
rice2 80 1.00 
rug1 23 37 0.46 
sand1 74 0.93 
scarf1 80 1.00 
scarf2 80 1.00 
screen1 74 0.93 
seat1 79 0.99 
seat2 72 0.90 
sesameseeds1 71 0.89 
stone1 79 0.99 
stone2 69 0.86 
stone3 65 0.81 
stoneslab1 74 0.93 
wall1 17 11 50 0.63 
Average  0.91 
Close Modal

or Create an Account

Close Modal
Close Modal