Gold-Price Forecasting Method Using Long Short-Term Memory and the Association Rule

Since gold prices influence international economic and monetary systems, numerous studies have been conducted to forecast prices. Nonetheless, employing the linear relationship method usually fail explain change in pattern of price. This study introduces a new paradigm that incorporates association rules long short-term memory (LSTM) as nonlinear-based method. For simulation, proposed was analyzed with data from Yahoo Finance January 2010 December 2020. The rule used choose features relevant spot (GS) US Dollar Index (DXY). LSTM price range hyperparameter settings. simulation results showed method—the GS DXY, or LSTM-GS-DXY—resulted low mean absolute percentage error (MAPE) metrics. In addition, LSTM-GS-DXY system outperformed simple moving average (SMA), weight (WMA), exponential (EMA), auto-regressive integrated (ARIMA).