Prediction of gold and silver stock price using ensemble models

Gold price prediction is a complex problem due to its non-linearity and dynamic time series behavior, constrained with many factors like economic, financial etc. Due to its high degree of monetary rewards and understanding the hidden pattern behind stock prediction researchers have proposed many statistical and machine learning algorithms for stock prediction. In this paper we examine different ensemble models for determining the future momentum of the gold and silver stock price, whether it will increase or decrease for the following relative to current days stock price. Using stacking approach we got significant accuracy of 85 % for predicting gold stock and 79 % for silver stock using a hybrid bagging ensemble.