Silver price prediction analysis using hybrid deep neural network model based on commodity correlation

Predicting the value of raw materials is not an easy job to do. A good silver price prediction can be made based on the historical price of silver and the prices of those highly correlated commodities. Historical prices for silver and 19 other commodities were taken from Metatrader 5 and Yahoo! data sources. Finance. The correlation with Silver was evaluated to select those products with high correlation [positive or negative] with Silver. Historical prices of highly correlated products were the input to the neural network to evaluate performance based on root mean square error [RMSE]. The silver price prediction obtained on highly correlated commodities is better than that obtained using only silver as input, it is also more accurate than that obtained based on uncorrelated commodities.