Hamid, Hadi Dwi Abdullah (2019) Diagnosis Penyakit Tanaman Cabai Menggunakan Metode Modified K-Nearest Neighbor (MKNN). Sarjana thesis, Universitas Brawijaya.
Abstract
Cabai merah merupakan salah satu jenis sayuran yang cukup penting di Indonesia, baik sebagai komoditas konsumsi di dalam negeri maupun sebagai ekspor. Cabai merah memiliki nilai gizi tinggi, juga mempunyai nilai ekonomi tinggi. Namun produktivitas cabai merah nasional rendah yaitu 7,34 ton/ha, padahal potensinya mampu mencapai 12 ton/ha. Dalam suatu periode tanam, cabai bisa dipanen beberapa kali bila musim dan perawatannya baik dapat 15-17 kali, namun umumnya hanya 10-12 kali. Rendahnya produktivitas bisa disebabkan oleh berbagai faktor, yaitu mutu benih kurang baik, tingkat kesuburan tanah semakin menurun, teknik budidaya kurang baik, serta permasalahan hama dan penyakit tanaman. Untuk menangani hal tersebut diperlukan teknologi dengan menerapkan metode klasifikasi yaitu Modified K-Nearest Neighbor (MKNN). MKNN adalah pengembangan dari metode KNN yang dirancang untuk mengatasi kelemahan dari jarak data dengan weight pada KNN. Metode tersebut mempelajari berdasarkan 18 gejala penyakit dengan proses perhitungan jarak euclidean, perhitungan validitas dan perhitungan wighted voting yang menghasilkan penetapan kelas klasifikasi berdasarkan nilai K yang ditentukan. Hasil pengujian menggunakan K=5 mendapatkan akurasi sebesar 94%, kemudian K=8 akurasi sebesar 92%, K=11 akurasi sebesar 88% dan pengujian K=14 menghasilkan akurasi sebesar 88%. Berdasarkan hasil yang didapat, metode Modified K-Nearest Neighbor (MKNN) menunjukkan akurasi yang baik untuk melakukan klasifikasi penyakit cabai.
English Abstract
Red chili is one of the most important vegetables in Indonesia, whether it is as a commodity that is consumed domestically and as an export commodity. As vegetables, beside red chili has a high nutritional value, it also has a high economic value. However, the productivity of national red chilli is still very low at 7.34 tons / ha, whereas the actual yield can potentially reach 12 tons / ha. In a planting period, chili can be harvested several times. If the season and the treatment is very good, chili can be harvested 15-17 times but generally, it can be harvested only 10-12 times. The low productivity of chili can be caused by a variety of factors, including poor quality of chili seeds, decreasing soil fertility, bad implementation of cultivation techniques, plant pest and disease problems. In order to handle this, technology is needed by applying one of the classification methods, namely Modified K-Nearest Neighbor (MKNN). Modified K-Nearest Neighbor (MKNN) is the development of the KNN method that has been designed to overcome the weaknesses of the distance between data and weight in the KNN. The method analysis is based on 18 symptoms of the disease with the process of calculating euclidean distance, calculating the validity and calculation of wighted voting which result in the determination of the classification class based on the specified K value. The test results showed that when using the value K = 5 produces an accuracy of 94%, then K = 8 produces an accuracy of 92%, K = 11 produces an accuracy of 88% and testing K = 14 produces an accuracy of 88%. Based on the results obtained, the Modified K-Nearest Neighbor (MKNN) method showed good accuracy for classifying chili disease.
| Item Type: | Thesis (Sarjana) |
|---|---|
| Identification Number: | SKR/FTIK/2018/1051/051901080 |
| Uncontrolled Keywords: | Modified K-Nearest Neighbor (MKNN), Klasifikasi, Penyakit, Cabai-Modified K-Nearest Neigbor (MKNN), Classification, Disease, Chili. |
| Subjects: | 600 Technology (Applied sciences) > 632 Plant injuries, diseases, pests > 632.3 Diseases / Plant diseases |
| Divisions: | Fakultas Ilmu Komputer > Teknik Informatika |
| Depositing User: | soegeng sugeng |
| Date Deposited: | 23 Oct 2019 04:11 |
| Last Modified: | 19 Oct 2021 08:11 |
| URI: | http://repository.ub.ac.id/id/eprint/167119 |
Preview |
Text
Hadi Dwi Abdullah Hamid.pdf Download (4MB) | Preview |
Actions (login required)
 |
View Item |