Designing Fuzzy Time Series Model and Its Application to Forecasting Inflation Rate

Agus Maman, Abadi and Subanar and Widodo and Samsubar, Saleh (2008) Designing Fuzzy Time Series Model and Its Application to Forecasting Inflation Rate. (Unpublished)

Full text not available from this repository.


Fuzzy time series is a dynamic process with linguistic values as its observations. Modelling fuzzy time series developed by some researchers used the discrete membership functions and table lookup scheme from training data. Table lookup scheme is a simple method that can be used to overcome the conflicting rule by determining each rule degree. The weakness of fuzzy time series model based on table look up scheme is that the fuzzy relations may not be complete so the fuzzy relations can not cover all values in the domain. This paper presents new method to modelling fuzzy time series combining table lookup scheme and singular value decomposition methods which use continuous membership function. Table lookup scheme is used to construct fuzzy relation from training data and then singular value decomposition of firing strength matrix is used to reduce fuzzy relations. Furthermore, this method is applied to forecast inflation rate in Indonesia based on six-factors one-order fuzzy time series. This result is compared with neural network method and the proposed method gets a higher forecasting accuracy rate than the neural network method. Key words. fuzzy time series, fuzzy rule, table lookup scheme, firing strength matrix, singular value decomposition, inflation rate.

Item Type: Article
Subjects: Matematika dan Ilmu Pengetahuan Alam > Matematika
Divisions: Fakultas Matematika dan Ilmu Pengetahuan Alam (FMIPA) > Pendidikan Matematika > Matematika
Depositing User: Agus Maman Abadi,S.Si.,M.Si.
Date Deposited: 20 Jul 2012 05:58
Last Modified: 20 Jul 2012 05:58

Actions (login required)

View Item View Item