Model Prediksi Harga Komoditas Cabai Merah Besar dan Keriting dengan Metode Long Short Term Memory

Abstract

Data Badan Pusat Statistik menandai tingginya tingkat produksi dan konsumsi cabai merah per kapita pada beberapa provinsi di Indonesia. Berdasarkan monitor Pusat Informasi Harga Pangan Strategis dari Mei 2018 hingga Mei 2021 harga cabai merah mengalami fluktuasi di 34 provinsi, salah satunya yaitu provinsi Jawa Barat. Penelitian ini bertujuan untuk membangun pemodelan prediksi harga cabai merah besar dan keriting di provinsi Jawa Barat dengan metode Long Short Term Memory (LSTM). Model prediksi harga komoditas cabai merah besar dan keriting menggunakan LSTM telah berhasil dibentuk dan dinilai cukup representatif untuk memprediksi harga di pasar tradisional dan pasar modern provinsi Jawa Barat. Hasil model prediksi terbaik untuk harga cabai merah besar dan keriting di pasar tradisional diperoleh nilai RMSE terkecil pada data uji sebesar 2,57% dan 2,07%. Sedangkan, hasil model prediksi harga terbaik di pasar modern diperoleh nilai RMSE terkecil pada data uji sebesar 2,11% dan 2,17%. Berdasarkan nilai RMSE yang diperoleh, pembentukan model menggunakan LSTM sudah lebih baik dari metode penelitian sebelumnya dan menunjukan bahwa variasi nilai yang dihasilkan pada model mendekati variasi nilai aktualnya.

Data from the Central Statistics Agency indicates the high level of production and consumption of red chili per capita in several provinces in Indonesia. Based on the monitoring of the Strategic Food Price Information Center from May 2018 to May 2021 the price of red chili fluctuated in 34 provinces, one of which was West Java province. The quantity of chili supply and the amount of demand imbalance needed by consumers causes price fluctuated. This study aims to build a predictive modeling of the price of large and curly red chilies in West Java Province using the Long Short Term Memory method. The red chili price prediction model using LSTM has been successfully formed and is considered representative enough to predict prices in traditional markets and modern markets in West Java Province. The best prediction model for the price of large and curly red chilies in traditional markets obtained the smallest RMSE values on the test data of 2.57% and 2.07%, respectively. Meanwhile, the best price prediction model in the modern market obtained the smallest RMSE values on the test data of 2.11% and 2.17%, respectively. Based on the RMSE value obtained, the model is better than the previous research method and shows that the variation in the value produced by a model is close to the variation in the actual value.