Forecasting Model on Fuzzy Markov Chain

Forecasting plays a major role in various fields to predict the future events. Song and Chissom [11] were the pioneers of studying such problem and have proposed fuzzy time series model. This paper attempts to improve the accuracy of forecasting by Markov chain and classification states. The computation is carried out using fuzzy sets and a procedure is given to choose appropriate model for a given time series. Sullivan and Woodall [14] worked on comparison of fuzzy time series and Markov model forecasting using university enrollment of Alabama. The forecasting accuracy is further improved in the present work. The method is applied to the data set related to the Enrollment data on university of Alabama. The results are verified using residual variance and average forecasting error by comparing with existing methods.