Neural Network Based Mathematical Models for Gold Price Prediction

Shetty, Dasharathraj K and Annappa, K and Aithal, Prakash K and Melenta, Vrishika (2014) Neural Network Based Mathematical Models for Gold Price Prediction. International Journal of Advanced Computer Communications and Control, 2 (4). pp. 134-138.

[img] PDF
INTERNATIONAL JOURNAL OF ADVANCED COMPUTER COMMUNICATIONS & CONTROL.pdf - Published Version
Restricted to Registered users only

Download (463kB) | Request a copy

Abstract

Gold is an old, rare and precious metal and is used as a global currency, a commodity, an investment and as a metal for ornaments. It has been used for coinage, jewellery, and other arts historically. There are several factors affecting the price of Gold prominent ones being Population, Inflation, Dollar Rate, rate of Bank Interest and Stock etc. This research project proposes a mathematical model that predicts the average annual price of Gold. Five mathematical models viz., Decision Tree,Regression, Radial Basis with exact fit, Radial basis with fewer Neurons (equal to number of input attributes) and Multilayer Feed Forward Network were experimented, and the results were compared with the last 129 (1883 to 2011) years actual average annual price of Gold. Both Multilayer Feed Forward Network with Back-Propagation and Radial Basis Function with fewer Neurons predicted best results with 99.504% accuracy. Following cases of input parameters were tested viz., (i) Population and Inflation (ii) Population alone

Item Type: Article
Uncontrolled Keywords: Gold Price Prediction, Artificial Neural Network, Radial Basis, Multilayer Feed Forward Network, Decision Tree
Subjects: Engineering > MIT Manipal > Computer Science and Engineering
Depositing User: MIT Library
Date Deposited: 16 May 2015 07:09
Last Modified: 16 May 2015 07:09
URI: http://eprints.manipal.edu/id/eprint/142657

Actions (login required)

View Item View Item