MODELING OF CAPACITY UTILIZATION RATIO WITH FUZZY TIME SERIES BASED ON MARKOV TRANSITION MATRIX

Abstract

Modeling of time series with fuzzy logic has found an increasinglyexpanding usage in recent years. One of the most important reasons for thisis that fuzzy logic approach doesn't require assumptions needed by the typicaltime series. Inclusion of the some weightings and probability calculations at theforecast stage into the ...rst studies starting through the modeling of time serieswith fuzzy logic resulted in further improvement of the forecast quality. Tsaur(2011) achieved better forecasted results by including the Markov transitionprobabilities matrix. Fuzzy time series is also an approach which can be *exiblyused in various model structures as it easily overcomes the di¢ culties causedby the model structure - linear or non-linear form. In this study, Markovmethod of Tsaur is applied on the monthly capacity utilization ratio (CUR)of Turkey which has a non-linear structure and free of seasonality belongingto the period between 2007-2015. In this sense, the results are compared tothe results of SETAR model and it's seen that Tsaur's approach has providedbetter results compared to the forecasts of typical time series