MESIN PEMBELAJARAN ENSEMBLE UNTUK IDENTIFIKASI VARIETAS PADI
| dc.contributor.author | Ikhlas, Ariza | |
| dc.contributor.author | Abdullah | |
| dc.contributor.author | Prasetyo, Dwi Yuli | |
| dc.date.accessioned | 2024-10-24T01:57:45Z | |
| dc.date.available | 2024-10-24T01:57:45Z | |
| dc.date.issued | 2020-06-15 | |
| dc.description | Setiap varietas padi memiliki karakter tertentu dengan anjuran tanam berbeda. Petani umumnya kesulitan memilih varietas padi yang cocok untuk ditanam di lahan mereka karena kurangnya kemampuan identifikasi. Algoritma klasifikasi merupakan solusi mengatasi masalah ini karena mampu mengidentifikasi varietas padi melalui citra digital. Tujuan penelitian ini adalah menerapkan dan mengevaluasi beberapa algoritma klasifikasi untuk mengidentifikasi varietas padi menggunakan fitur warna dan tekstur. Penelitian dilakukan di kabupaten Indagiri Hilir Riau pada tahun 2018. Mesin pembelajaran dibangun dengan cara menggabungkan beberapa algoritma klasifikasi (classifier), yaitu Support Vector Machine, k-Nearest Neighbors, Logistic Regression, dan Decision Tree. Varietas yang diteliti adalah IR42, Inpara-9. dan Batang Piaman. Berdasarkan tingkat ketelitian masing-masing algoritma, k-Nearest Neighbors memberikan hasil lebih baik dibanding algoritma lainnya, baik dengan maupun tanpa normalisasi data. Terdapat enam sampel Inpara-9 yang diprediksi benar (true positive) dan lima sampel diprediksi salah (false positive). Pada varietas Batang Piaman terdapat delapan sampel yang diprediksi benar (true positive). Pada IR42 terdapat lima sampel yang diprediksi benar. | |
| dc.description.abstract | Each rice variety has certain characteristics with different planting recommendations. Farmers generally find it difficult to select rice varieties suitable for planting on their land due to a lack of identification capabilities. The classification algorithm is a solution to this problem because it is able to identify rice varieties through digital images. The purpose of this study was to apply and evaluate several classification algorithms to identify rice varieties using color and texture features. The research was conducted in Indagiri Hilir Riau district in 2018. Machine learning was built by combining several classification algorithms (classifier), namely Support Vector Machine, k-Nearest Neighbors, Logistic Regression, and Decision Tree. The varietes studied were varieties IR42, Inpara-9. and Batang Piaman. Based on the level of accuracy of each algorithm, k-Nearest Neighbors gives better results than other algorithms, both with and without normalization data. There were six samples of Inpara-9 that were predicted to be true (true positive) and five samples were redicted to be false (false positive). In the Batang Piaman, there were eight samples that were predicted to be true (true positive). In IR42, there were five samples that were predicted to be true. | |
| dc.identifier.issn | 0852-1743 | |
| dc.identifier.uri | https://repository.pertanian.go.id/handle/123456789/23750 | |
| dc.language.iso | id | |
| dc.publisher | Sekretariat Badan Penelitian dan Pengembangan Pertanian | |
| dc.relation.ispartofseries | Vol. 29; No. 2 | |
| dc.title | MESIN PEMBELAJARAN ENSEMBLE UNTUK IDENTIFIKASI VARIETAS PADI | |
| dc.title.alternative | Ensemble Machine Learning for Rice Varieties Identification | |
| dc.type | Article |