Research Article

Novel Approach to Classify Plants Based on Metabolite-Content Similarity

Table 4

Resulting confusion matrix from support vector machine (SVM) algorithm. 162 plants are labeled as edible (E), medicinal (M), timber (T), landscaping (L), and poisonous (P), and SVM model was constructed to classify them.

ā€‰METLPRecognition rate [%]

M810000100
E14700097.9
T6080057.2
L8105035.7
P410000

Total: 162 plants. Accuracy: 87.0%.