Implementasi Algoritma LSTM untuk Prediksi Harga Cabai Merah Keriting di Yogyakarta

Abstract

Curly red chili is one of the horticultural crops with high economic value. Curly red chili can cause inflation because the price fluctuates greatly and tends to rise. The purpose of the research is to predict future prices quickly and accurately so that the government and consumers can take preventive action against existing problems. This research will use the LSTM algorithm and a dataset of curly red chili prices in Yogyakarta from October 9, 2017 to October 27, 2023. Based on the testing of this research, the best results obtained are by using a data division ratio of 70%: 30%, epoch 150, batch size 48, learning rate 0.001, number of neurons 30, activation function ReLU, and the optimization function Adam which produces a MAPE value of 3.6995%, and an accuracy of 96.3005%. It is hoped that this system can help related parties get accurate curly red chili price predictions.