PERBANDINGAN LSTM DAN ELM DALAM MEMPREDIKSI HARGA PANGAN KOTA TASIKMALAYA

Humans have needs that must be met, one of which is the need for food, but food prices often change. Factors that affect price changes occur because the amount of demand is high while the supply is small. Making predictions about price changes will be very helpful to get an idea of the pattern of price changes. Therefore making predictions from price patterns is useful for providing information to the public. Predictions regarding price changes can be made using many methods. Long Short-Term Memory (LSTM) and Extreme Learning Machine (ELM) are several methods that can be used to predict time series data, these two methods can provide an overview of the predictions made. The results of the study show that both algorithms have good results in terms of the the evaluation value. The evaluation results showed no significant difference between the two algorithms. The evaluation value of the rice commodity showed that ELM tended to be better with MAE values of 6,721, MAPE 0.061%, MSE 115,281, RMSE 10,737 and CV 3,699%, while LSTM with MAE 31,707, MAPE 0.286%, MSE 1927.633, RMSE 43.905 and CV 3.655%. However, for other commodities, LSTM can produce a better evaluation value.