ACADSTAFF UGM

CREATION
Title : Sarima-egarch model to reduce heteroscedasticity effects in network traffic forecasting
Author :

Isna Alfi Bustoni, S.T., M.Eng. (2)

Date : 2017
Keyword : Forecasting, Bandwidth, Heteroscedasticity, SARIMA, EGARCH Forecasting, Bandwidth, Heteroscedasticity, SARIMA, EGARCH
Abstract : Difference needs in bandwidth allocations have not been accommodated by static bandwidth allocations that leads to ineffective bandwidth use. There are several previous researches about bandwidth allocations which have been conducted, such as the use of Seasonal Autoregressive Integrated Moving Average (SARIMA) method. However, SARIMA method is not able to overcome various kinds of error problems or heteroscedasticity. Therefore, this research proposes the application of SARIMA-EGARCH (Exponential Generalized Autoregressive Conditional Heteroscedastic) method to generate the more accurate model that is able to overcome heteroscedasticity on the needs of bandwidth forecasting. In addition, this research compares the result of SARIMA to SARIMA-EGARCH examinations. It shows that SARIMA (1,0,1) (3,1,1)7 has 11,38% Mean Absolute Percentage Error (MAPE) and SARIMA-EGARCH (1,0,1)(3,1,1)7(1,1) has only 9,20%. The comparison shows that applying EGARCH increase the accuracy to 19,15%.
Group of Knowledge : Ilmu Komputer
Original Language : English
Level : Internasional
Status :
Published
Document
No Title Document Type Action