Wood is a commodity that is usually used for various purposes. Growing demands for timber have led some humans to commit fraudulent and unlawful moves. One of them is through illegal logging of wood species that are protected by the state. The Ministry of Environment and Forestry apart from forest rangers has several problems in classifying wood species. Then technology is used to overcome these problems through the use of machine learning. One suitable algorithm for classifying is K-NN. There are five types of information wood used, specifically marine resin, teak, kruing, meranti, and ironwood. The total wood photos are 1,300 with a complete dataset of 250 images for each type of wood and 10 images for classification testing. The trial was carried out by finding the most optimal K, ie K = 3, after that it was recorded with a confusion matrix. The results obtained are 76% accuracy, 78.8% recall, 76% precision, and 77.37% F1-Score. The higher the value of K, the greater the number of classifications and the lower the accuracy. The higher the value of K, the more types of classifications you want to test, and the less accurate the percentages in the classification process.
Copyrights © 2022