Predict the Price of Gold Based on Machine Learning Techniques

Han-chao ZHU, Dong WANG

Abstract


Due to the unique hedging and investment functions of the precious metals market, the precious metals represented by gold play a decisive role in the international financial market. The forecast of the gold price has drawn great attention from investors, government departments and researchers all over the world. This paper made an innovation in predicting the price of precious metals, bidding farewell to the traditional statistical models. Taking the gold as an example, a machine learning method to completely predict the trend of the gold price and proved its reliability by experiments. The experimental results show that our proposed method has a significant advantage in forecasting gold price movements and provides a more reliable and effective result than any other methods proposed before. The main contributions of this paper are as follows: 1. The association rules algorithm is introduced, adjusted and modified to solve the problem of factor mining in predictive models, so that the computer can mine the most time-efficient influencing factors according to the latest data set. G,M(1,1) model is proposed to successfully predict the gold price trend combining with the mining of above factors. Since our proposed model can easily be extended to forecasts of other economic targets such as stock index and other commodity prices, we expect our method to be scaled to more areas of forecasting.

Keywords


Machine learning, Grey model, Association rules mining, Prediction


DOI
10.12783/dtcse/mmsta2017/19700

Full Text:

PDF

Refbacks

  • There are currently no refbacks.