Implementing Method of Ensemble Empirical Mode Decomposition And Recurrent Neural Network For Gold Price Forecasting

Abstract

Gold becomes one of long-term investment options and is used as a value protection against inflation or declining other assets. These gold price fluctuations tend to be nonlinear and uncertain. Most researchers and business practitioners fail to produce consistent pricing analyses, due to the complexity of the dynamic and volatile gold market. One method that can accommodate gold price fluctuations is using Ensemble Empirical Mode Decomposition (EEMD). Furthermore, the results of the gold price analysis can be used in forecasting. Forecasting fluctuations in gold prices are needed by importers, investors, and society to reduce risks and to help in making decision. The forecasting which has been done is the integration between EEMD and Feed-forward Neural Network (FNN) with good forecasting results. However, the use of FNN is less flexible for the use of free parameters, such as the type of activation function, initial initialization, number of input neurons, and output neurons. The setting of flexible free parameters can affect the performance of neural networks and improve forecasting accuracy. One way to overcome the weaknesses of FNN in the use of free parameters is, it can use the Recurrent Neural Network (RNN). The trial in this study is using monthly data of world gold price. The results proves that the performance of EEMD-RNN method forecasting is better than EEMD-FNN.

Authors and Affiliations

Sri Herawati, Firmansyah Adiputra, M. Latif, Aeri Rachmad

Keywords

Related Articles

Wireless Body Area Networks

In today’s environment health care monitoring is very important area of research. Wireless Sensor Networks (WSNs) can also be used in various health care applications. Wireless .Body Area Networks (WBANs) are getting dev...

Effect of Gradation and Plasticity on Compaction Characteristics of Red Gravel Soils

Red gravel soils are widely distributed over the north coastal districts of Andhra Pradesh. These soils are highly desirable for construction activities due to their inherent advantages. To use these soils in various Geo...

Mass Transfer Effects on Stokes Problem for an Infinite Vertical Plate in a Rotating Fluid.

An exact solution to the unsteady free convection flow of viscous incompressible fluid, in the presence of foreign mass, past an impulsively started infinite vertical isothermal plate in a rotating fluid, has been derive...

Nanoparticle Evaluation of Synthetic Palmitoyl-CKKHH As Transfection Reagent for Non-Viral Gene Delivery Vehicle

Non-viral delivery systems are relatively safe but inefficient in their current form. The main obstacle in using non-viral gene delivery system approach is to transport the gene of interest in cytoplasm and subsequently...

Stabilization of Sewage Sludge Using Cow Dung, Poultry Manure and Horse Dung by Vermicomposting

Toxic concentrations of heavy metals have been reported to be present in STPs as many industries dispose their waste directly in domestic sewers without prior treatment and thus they are carried to the sludge generated d...

Download PDF file
  • EP ID EP392641
  • DOI 10.9790/9622-0711013943.
  • Views 74
  • Downloads 0

How To Cite

Sri Herawati, Firmansyah Adiputra, M. Latif, Aeri Rachmad (2017). Implementing Method of Ensemble Empirical Mode Decomposition And Recurrent Neural Network For Gold Price Forecasting. International Journal of engineering Research and Applications, 7(11), 39-43. https://europub.co.uk/articles/-A-392641