Pengembangan Decision Support System Berbasis Hybrid LSTM Dan Rule-Based Untuk Optimalisasi Waktu Tanam Padi Di Sumatera Utara

Sipayung, Retha Novianty (2025) Pengembangan Decision Support System Berbasis Hybrid LSTM Dan Rule-Based Untuk Optimalisasi Waktu Tanam Padi Di Sumatera Utara. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5026211028-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (9MB) | Request a copy

Abstract

Beras sebagai makanan pokok masyarakat Indonesia memiliki peran strategis dalam ketahanan pangan nasional. Optimalisasi produksi masih terkendala oleh ketepatan waktu tanam yang bergantung pada pola curah hujan. Tugas akhir ini mengusulkan pengembangan Decision Support System (DSS) berbasis hybrid yang mengintegrasikan Long Short-Term Memory (LSTM) dengan Rule-Based System untuk optimalisasi waktu tanam padi di Sumatera Utara. Metodologi tugas akhir mencakup pengumpulan data harian meteorologi periode 2014-2024, data harga beras 2021-2024, dan data serangan hama 2020-2024. Setelah preprocessing, model LSTM dikembangkan untuk prediksi curah hujan dan klasifikasi hari kering/basah, dengan konfigurasi optimal diperoleh melalui eksperimen grid search. Hasil prediksi LSTM kemudian diintegrasikan ke dalam Rule-Based System yang menghasilkan rekomendasi waktu tanam dalam format kode warna. Evaluasi sistem menunjukkan model prediksi curah hujan memiliki MAE sebesar 12,93 mm dan R² sebesar 0,1677, model klasifikasi hari kering mencapai akurasi 70,07% dengan F1-score 64,4%, dan model prediksi harga beras menunjukkan MAE 0,07 ribu rupiah dengan R² 0,6762. Rule-Based System menunjukkan akurasi 69,28% saat dibandingkan dengan label aktual dari data BMKG. Sistem ini diimplementasikan dalam antarmuka interaktif Streamlit yang menampilkan rekomendasi waktu tanam beserta informasi pendukung. Tugas akhir ini berkontribusi pada pengembangan sistem pendukung keputusan pertanian yang menggabungkan kemampuan prediktif machine learning dengan transparansi Rule-Based System, memberikan rekomendasi waktu tanam yang komprehensif bagi petani di Sumatera Utara.
====================================================================================================================================
Rice, as the staple food of Indonesian society, plays a strategic role in national food security. Production optimization is still constrained by planting timing, which depends heavily on Rainfall patterns. This thesis proposes the development of a hybrid Decision Support System (DSS) that integrates Long Short-Term Memory (LSTM) with a Rule-Based System to optimize rice planting times in North Sumatra. The methodology includes collecting daily meteorological data from 2014-2024, rice price data from 2021-2024, and pest attack data from 2020-2024. After preprocessing, LSTM models were developed for Rainfall prediction and dry/wet day classification, with optimal configurations obtained through grid search experiments. The LSTM prediction results were then integrated into a Rule-Based System that generates planting time recommendations in a color-coded format. System evaluation shows that the Rainfall prediction model has an MAE of 12.93 mm and R² of 0.1677, the dry day classification model achieves 70.07% accuracy with an F1-score of 64.4%, and the rice price prediction model shows an MAE of 0.07 thousand rupiah with R² of 0.6762. The Rule-Based System demonstrates 68.28% accuracy when compared with actual labels derived from BMKG data. The system is implemented in an interactive Streamlit interface that displays planting time recommendations along with supporting information. This thesis contributes to the development of agricultural decision support systems that combine the predictive power of machine learning with the transparency of Rule-Based Systems, providing comprehensive planting time recommendations for farmers in North Sumatra.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Decision Support Systems, LSTM, Rule-Based System, Prediksi Curah Hujan, Waktu Tanam Padi, Rainfall Prediction, Rice Planting Time
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > QC Physics > QC866.5 Climatology--Forecasting. T Technology > T Technology (General) > T174 Technological forecasting T Technology > T Technology (General) > T58.62 Decision support systems
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information System > 57201-(S1) Undergraduate Thesis
Depositing User:	Retha Novianty Sipayung
Date Deposited:	23 Jul 2025 02:26
Last Modified:	23 Jul 2025 02:26
URI:	http://repository.its.ac.id/id/eprint/120696

Actions (login required)

View Item