Prediction of Gold Price using Deep Learning

Abstract

Gold has been the original value and medium of exchange from many centuries. Gold price prediction in India in 2021 according to the last previous rededication is Rs 60,300. By predicting gold price investors will have an idea about when to buy or sell the commodity. Gold price is directly linked to the country's currency and hence affects the stock price. It is seen that the decrease in stock price increases the gold price. This paper is mainly aiming on the early prediction model using Long Short Term Memory (LSTM) and its variance. The proposed model is compared with the widely used linear regression algorithm. It is observed that the Mean Absolute Percentage Error (MAPE) value of Linear Regression, Vanilla, and Stacked LSTM is 10.94, 2.649, and 2.5009 respectively. From the values of MAPE it is observed that LSTM has outperformed compared to Linear regression. Hence early prediction of gold price helps investors to invest or sell the gold at the best possible price.