الفهرس | Only 14 pages are availabe for public view |
Abstract The aim of this thesis is a pursuit to model the Egyptian stock exchange indexes using ARIMA process. This study assumes the Egyptian stock exchange index as a univariate stochastic process that only depends on its historical observed values. Box-Jenkins’ autoregressive integrated moving average (ARIMA) modelling process is applied to study the characteristics of the studies series and to calculate suitable forecasts. Then, the calculated forecasts are compared to those found using artificial neural network (ANN) techniques. It was found that ARIMA models can be used in some cases to provide satisfactory short-term forecasts. It was also found that major rapid changes in the observed values produce larger errors. Hence, using longer history does not always produce more accurate results when compared to using shorter historical data sample. |