Prediction of Gold Prices Using Artificial Neural Networks

Abstract

Gold, which has fulfilled the functions that money has in the past years, is now mostly used as a means of saving. Gold has been an investment instrument for Turkish society for many years. In addition, it is perceived as the most reliable means of investment by people in times of crisis and war. It is very important to estimate gold prices for investors who want to earn high profits from their investments. In this study, as input (independent variables), the data set consisting of Brent oil price, USD price, BIST100 index, Central Bank of the Republic of Turkey weekly interest rate, silver and copper prices was applied to the price of gold in ounce by applying Multi-Layer Perceptron Neural Network (MLPNN) is intended to be estimated. As the data set, weekly price and index values were used between January 2010 and December 2016 period. As a result of the study, suggestions were made about the effectiveness of gold on the input variables created by using MLPNN.