PENERAPAN METODE LONG SHORT TERM MEMORY UNTUK MEMPREDIKSI HARGA BERAS DI INDONESIA

THE APPLICATION OF LONG SHORT TERM MEMORY METHOD FOR PREDICTING RICE PRICES IN INDONESIA

Beras merupakan makanan pokok masyarakat di Indonesia. Berdasarkan data Sistem Pemantauan Pasar dan Kebutuhan Pokok (SP2KP) Kementrian Perdagangan, harga beras di Indonesia terus mengalami kenaikan sejak Agustus 2022 (Arinal & Azhari, 2023), hal ini berdampak pada tingkat kemiskinan, tingkat inflasi, dan stabilitas ekonomi. Oleh karena itu, pemerintah membutuhkan Solusi agar dapat menghindari dampak dari naik turunnya harga beras di Indonesia. Metode Long Short Term Memory (LSTM) memiliki kemampuan dalam menangani masalah ketergantungan jarak jauh pada data berurutan, sehingga cocok digunakan untuk memprediksi harga beras. Kombinasi parameter yang digunakan adalah metode normalisasi data, pembagian data, layer, batch size, epoch, dan learning rate. Hasil evaluasi model menggunakan MAPE (Mean Absolute Percentage Error) menunjukkan bahwa LSTM dapat memprediksi harga beras dengan akurasi 98.57% dan nilai MAPE 1.43%. Hasil prediksi terbaik didapatkan dengan menggunakan parameter metode normalisasi data StandardScaler, pembagian data latih 80%, data validasi 10%, data uji 10%, layer 2, batch size 4, epoch 40, dan learning rate 0.01, dengan pembagian data secara acak. Performa metode LSTM terbukti memiliki performa yang lebih baik. Untuk mengetahui performa LSTM, dilakukan perbandingan dengan metode ARIMA (Autoregressive Integrated Moving Average) dan RNN (Recurrent Neural Network). Nilai MAPE terbaik dari LSTM 1.43%, ARIMA 10.41%, dan RNN 3.53%. Adapun hasil akurasi terbaik LSTM ialah 98.57%, ARIMA 89.59%, dan RNN 96.47%.

KATA KUNCI: Long Short Term Memory, Mean Absolute Percentage Error, Autoregressive Integrated Moving Average, Recurrent Neural Network, prediksi, beras

Rice is the staple food for the Indonesian population. According to data from the Market Monitoring and Basic Needs System (SP2KP) of the Ministry of Trade, rice prices in Indonesia have been consistently rising since August 2022 (Arinal & Azhari, 2023). This increase impacts poverty levels, inflation rates, and economic stability. Therefore, the government requires solutions to mitigate the effects of fluctuations in rice prices in Indonesia. The Long Short Term Memory (LSTM) method is capable of addressing long-term dependencies in sequential data, making it suitable for predicting rice prices. The parameter combinations used include data normalization methods, data splitting ratios, layers, batch size, epochs, and learning rates. The evaluation results of the model using MAPE (Mean Absolute Percentage Error) indicate that LSTM can predict rice prices with an accuracy of 98.57% and a MAPE value of 1.43%. The best prediction results were achieved using the following parameters: StandardScaler as the data normalization method, 80% training data, 10% validation data, 10% test data, 2 layers, batch size of 4, 40 epochs, and a learning rate of 0.01, with data split randomly. The performance of the LSTM method proved superior. To assess LSTM's performance, a comparison was conducted with the ARIMA (Autoregressive Integrated Moving Average) and RNN (Recurrent Neural Network) methods. The best MAPE values were 1.43% for LSTM, 10.41% for ARIMA, and 3.53% for RNN. Meanwhile, the best accuracy values were 98.57% for LSTM, 89.59% for ARIMA, and 96.47% for RNN.

KEYWORDS: Long Short Term Memory, Mean Absolute Percentage Error, Autoregressive Integrated Moving Average, Recurrent Neural Network, prediction, rice