Authors
Abstract
New Approaches in Forecasting Using
Neuro-Fuzzy Networks
(Case Study: The Crude Oil Price)
Mohammad Rahim Ramezanian1, Esmael Ramezanpour2,
Sayeed Hamed Pourbakhsh3
1- Assistant Professor. of Management, Guilan University, Guilan, Iran
2- Assistant Professor of Economy, Guilan University, Guilan, Iran
3- B.sc of Industrial Management, Guilan University, Guilan, Iran
Received: 17 /4/2010 Accept: 18/4/2011
Our world is a rapidly changing world. So it is very important for the survival of organizations to know what lies ahead in the future, how much demand is there for their products and for what price? We cannot afford big changes unless we are able to predict the future. The application of predictability science in management has been studied in this research. With the increasing progress of science, the use of new methods and application of new intelligent technologies have also increased. In this research, new methods and algorithms such as neural networks and fuzzy logic have been explained and the application of their combination in predictability has been studied. Various methods such as Moving Average Method, Weighted Moving Average, Exponential Smoothing, Double Exponential Smoothing, Linear Trend, Combined Functional Trend, and Exponential Process were used to make predictions. The results obtained from these methods were compared with the those obtained from the neru-fuzzy networks method using 6 error measurement criteria. It was found that the neru-fuzzy networks method yielded better results, and the correspondence of data (R2 coefficient) for the Neru-fuzzy networks was about 90 percent. The data used in this research were related to the Organization of the Petroleum Exporting Countries (OPEC) for the years 1970-2000.
Keywords
)