ARDIANSYAH, RANGGA (2025) KLASIFIKASI PRODUKSI KEDELAI BERBASIS DEEP LEARNING MENGGUNAKAN CATBOOST. S1 thesis, Universitas Mercu Buana Jakarta.
|
Text (HAL COVER)
01 COVER.pdf Download (648kB) | Preview |
|
![]() |
Text (BAB I)
02 BAB 1.pdf Restricted to Registered users only Download (161kB) |
|
![]() |
Text (BAB II)
03 BAB 2.pdf Restricted to Registered users only Download (193kB) |
|
![]() |
Text (BAB III)
04 BAB 3.pdf Restricted to Registered users only Download (200kB) |
|
![]() |
Text (BAB IV)
05 BAB 4.pdf Restricted to Registered users only Download (383kB) |
|
![]() |
Text (BAB V)
06 BAB 5.pdf Restricted to Registered users only Download (143kB) |
|
![]() |
Text (DAFTAR PUSTAKA)
07 DAFTAR PUSTAKA.pdf Restricted to Registered users only Download (141kB) |
|
![]() |
Text (LAMPIRAN)
08 LAMPIRAN.pdf Restricted to Registered users only Download (347kB) |
Abstract
This study analyzes and predicts soybean yield categories (low, medium, or high) using two machine learning models: CatBoost Classifier and Neural Network. The research dataset focuses on 166,349 pre-processed soybean entries extracted from a total of 1 million crop yield records. Exploratory data analysis revealed significant class imbalance within the yield distribution. Both models demonstrated high classification performance, with CatBoost achieving an accuracy of 0.9374 and the Neural Network reaching 0.9348. However, further evaluation indicated that CatBoost was superior in handling minority classes, evidenced by its better recall for the low-yield category. Based on CatBoost's feature importance analysis, fertilizer use, rainfall, and irrigation were identified as the most crucial factors significantly impacting soybean yield. These findings offer valuable insights for agricultural practices, emphasizing the optimization of these key factors to enhance soybean productivity. Keywords: Soybean, Crop Yield, Classification, CatBoost, Neural Network, Feature Importance, Machine Learning CatBoost Classifier dan Neural Network. Dataset penelitian berfokus pada 166.349 entri data kedelai yang telah melalui proses pra-pemrosesan dari total 1 juta data panen. Hasil analisis data eksploratif menunjukkan adanya ketidakseimbangan kelas signifikan pada distribusi hasil panen. Kedua model menunjukkan performa klasifikasi yang tinggi, dengan CatBoost mencapai akurasi 0.9374 dan Neural Network 0.9348. Meskipun demikian, evaluasi lebih lanjut mengungkapkan bahwa CatBoost lebih unggul dalam menangani kelas minoritas, terbukti dari recall yang lebih baik untuk kategori panen rendah. Berdasarkan analisis feature importance dari model CatBoost, penggunaan pupuk, curah hujan, dan penggunaan irigasi merupakan faktor-faktor paling krusial yang secara signifikan memengaruhi hasil panen kedelai. Temuan ini memberikan wawasan penting bagi praktik pertanian untuk mengoptimalkan faktor-faktor kunci demi peningkatan produktivitas kedelai. Kata Kunci: Kedelai, Hasil Panen, Klasifikasi, CatBoost, Neural Network, Feature Importance, Machine Learning.
Actions (login required)
![]() |
View Item |