| Autor: |
Saputra, Rizal Amegia, Wasiyanti, Sri, Saefudin, Dede Firmansyah, Masripah, Siti, Utami, Lila Dini |
| Předmět: |
|
| Zdroj: |
AIP Conference Proceedings; 5/12/2023, Vol. 2714 Issue 1, p1-5, 5p |
| Abstrakt: |
Rice is a vital sector in many countries, especially in Indonesia. Indonesia is a rice producer of rice, and rice is the primary food for the Indonesian population. There are many risks of crop failure in rice farming, one of which is leaf disease. In this study, the KNN algorithm classifying rice leaf diseases. The dataset from public data published in the UCI Repository. Dataset consists of about 120 images with three rice leaf diseases, namely Bacterial leaf blight, Brown spot, Leaf smut. The image data processed into numeric numbers with the Gray Level Co-occurrence Matrix (GLCM) feature extraction technique. The KNN algorithm has weaknesses, one of which is that it must determine the maximum k value and is weak on non-informative variables. Particle Swarm Optimization (PSO) optimization algorithm is applied to find the best population size value to increase accuracy. The maximum k value is seven, and the population size 11. The highest accuracy value from the test results is 86.11%, and kappa 0.792. This result shows that the PSO-based KNN algorithm can classify rice leaf diseases with a kappa range on the Strength Of Agreement Good. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
