<![CDATA[In Banjar Regency, rice production faces significant challenges, including high crop failure rates and production variability across regions, which impact equitable food availability. This study aims to analyze the performance of various machine learning algorithms in predicting rice crop failures, a critical issue in food security. The research variables include factors such as weather, air humidity, soil conditions, agricultural variables, and tungro disease infestations. Several algorithms were tested, including Naive Bayes, Logistic Regression, Decision Tree, Random Forest, XGBoost, and others. Evaluation was conducted using cross-validation techniques with metrics such as accuracy, precision, recall, F1-Score, and ROC AUC. The results indicate that the Random Forest and XGBoost algorithms achieved the best performance, with accuracies of 77% and 70%, respectively. The study concludes that machine learning-based models can support better decision-making to mitigate crop failure risks. Furthermore, this research provides a foundation for the development of predictive models in the agricultural sector.Keywords: Harvest failure; Rice; Machine learning; Prediction; Food security AbstrakDi Kabupaten Banjar, produksi gabah menghadapi kendala signifikan, termasuk gagal panen yang tinggi dan variasi produksi antar wilayah, yang memengaruhi ketersediaan pangan merata. Penelitian ini bertujuan untuk menganalisis kinerja berbagai algoritma machine learning dalam memprediksi gagal panen gabah, yang merupakan permasalahan penting dalam ketahanan pangan. Variabel penelitian mencakup faktor-faktor seperti cuaca, kelembapan udara, kondisi tanah, variabel pertanian, dan serangan tungro. Beberapa algoritma yang diuji meliputi Naive Bayes, Logistic Regression, Decision Tree, Random Forest, XGBoost, dan lainnya. Evaluasi dilakukan menggunakan teknik cross-validation dengan metrik akurasi, precision, recall, F1-Score, dan ROC AUC. Hasil menunjukkan bahwa algoritma Random Forest dan XGBoost memberikan performa terbaik, dengan akurasi masing-masing sebesar 77% dan 70%. Kesimpulan penelitian ini menunjukkan bahwa model berbasis machine learning dapat digunakan untuk mendukung pengambilan keputusan yang lebih baik dalam mengurangi risiko gagal panen. Penelitian ini juga memberikan dasar untuk pengembangan model prediksi di sektor agrikultur.Kata kunci: Gagal panen; Gabah; Machine learning; Prediksi; Ketahanan pangan]]>
Copyrights © 2025
