Gold Price Prediction Using Machine Learning and Deep Learning

Volatility of gold prices has direct implications to the investors in Indonesia since it becomes very hard to identify the right time to invest in gold or when to sell the precious metal. Also, organizations that rely on gold experience difficulty in both cost and time control in relation to production since the price is constantly fluctuating. As a result, this paper seeks to respond to the above challenges by estimating and comparing models for gold price prediction in Indonesia. In contrast to those researches, which investigate only selected countries or do not evaluate the performance of the developed models, this paper compares LSTM and GRU deep learning models with a linear regression machine learning model. We assess their performance by the model fitness measures, which correspond to the R2 and charts The obtained results: LSTM 93%, GRU 96%, and the linear regression model 86% and got 0.3231 MAE, 0.1389 MSE, 0.3727 RMSE for Linear Regression, got 0.1500 MAE, 0.0397 MSE, 0.1992 RMSE for GRU. According to the evaluation results, we can conclude that the predicted model with the highest accuracy of predicting the gold price in Indonesia is the GRU model.