Efficient Prediction of Gold Prices Using Hybrid Deep Learning

Gold prices, in general, act counter to the market and are thus predictable to a certain degree based upon the fluctuations of other market entities. Neural networks have been applied to significant effect to difficult prediction problems and have achieved success in making predictions beyond traditional regression-based statistical models. Generally, such networks are hyper-specialized, and thus have effectiveness in a small subset of problems. This paper endeavors to take two models: RNN and CNN and hybridize them, creating a new model founded on their underlying logical theories and architectures. The primary goal in this paper is to show the effectiveness of the hybrid model in achieving a better gold price prediction that produces higher quality results at the cost of slightly increased complexity. Challenges and limitations of the proposed hybrid model are also addressed.