The effect of interval length in weighted subsethood fuzzy time series

Fuzzy time series has received increasing intentions since the first definition of fuzzy time series was introduced in 1993 by Song and Chissom. Then many studies proposed new fuzzy time series models. One of the steps in developing fuzzy time series model is partitioning the universe of discourse i...

Full description

Saved in:
Bibliographic Details
Main Authors: Mansor, Rosnalini, Zaini, Bahtiar Jamili, Othman, Mahmod, Kasim, Maznah Mat
Format: Conference Proceeding
Language:English
Subjects:
Forecasting techniques
Intervals
Partitioning
Time series
Universe
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Fuzzy time series has received increasing intentions since the first definition of fuzzy time series was introduced in 1993 by Song and Chissom. Then many studies proposed new fuzzy time series models. One of the steps in developing fuzzy time series model is partitioning the universe of discourse into different lengths of intervals. However, the performance of fuzzy time series model can be affected by interval length factor. Therefore, the objective of this study is to investigate the performance of Weighted Subsethood fuzzy time series model with different lengths of interval. By using the familiar data in fuzzy time series study, the historical enrollments of University of Alabama data, seven different number of interval cases were generated. The results show that partitioning the data into four intervals has the minimum forecasting error.
ISSN:0094-243X
1551-7616
DOI:10.1063/1.5041691