LAUSR

In this study, we develop a novel moving average forecasting approach based on fuzzy time series data set. The main objective of applying this moving average approach in develop method is to provide better results and enhance the accuracy in forecasted output by reducing the fluctuation in time series data set. The developed method is to define the universe of discourse and partition into equal length of intervals which is based on the average-length method. Further, triangular fuzzy sets are defined and obtain a membership grade of each moving average historical datum rather than actual datum of historical fuzzy time series data. Here, the fuzzification process of moving average historical data to their maximum membership grades obtained into corresponding triangular fuzzy sets. The general suitability of developed model has been examined by implementing in the forecast of student enrollments data at the University of Alabama. Further, the market price of State Bank of India share at Bombay Stock Exchange, India, has also been implemented in the forecast. The developed method of moving average fuzzy time series provides an improved forecasted output with least root mean square error and average forecasting errors which shows that our developed method is more superior than other existing models available in the literature based on fuzzy time series data.