Forecasting DiGi closing share prices using ARMA models

Tan, Mei Hoon (2008) Forecasting DiGi closing share prices using ARMA models. Universiti Malaysia Sabah. (Unpublished)

[img]
Preview
Text
ae0000002036.pdf

Download (2MB) | Preview

Abstract

The purpose of this research is to predict DiGi share market by using time series approach and ARMA model. In this study, the data being employed is from 3 January 2000 to 31 December 2007. The data range from 3 January 2000 to 30 August 2007 was used to form the ARMA models while the remaining of the data which is from 1 September 2007 to 31 December 2007 was used to validate the adequacy of the model formed. In this study, two series have been considered which are the series with unobserved data and the series with the unobserved data being estimated by using cubic spline approach. In this research, the order of the ARMA models has been set to p+q≤10. Therefore, there are 65 possible models for each series. These possible models will be eliminated to selected models and lastly the best model. In this research, four best models have been considered which are the best model from the series with the unobserved data and the best model from the series without unobserved data. Furthermore, there are another two alternative best models were taken into consideration in this research. However, at the end of this study, there is only one "best of the best model" being utilised in this study for future short term forecasting. The accuracy of the forecast was compared by using Mean Absolute Percentage Error(MAPE).

Item Type: Academic Exercise
Uncontrolled Keywords: DiGi, share market, time series, ARMA model, Mean Absolute Percentage Error
Subjects: Q Science > QA Mathematics
Divisions: SCHOOL > School of Science and Technology
Depositing User: MDM SITI AZIZAH IDRIS
Date Deposited: 18 Mar 2014 06:16
Last Modified: 13 Oct 2017 01:27
URI: http://eprints.ums.edu.my/id/eprint/8536

Actions (login required)

View Item View Item