Fuzzy forecasting methods and its applications in two different subjects

Humankind wants to know sequences of events before they haven't happened yet. Therefore, it tries to forecast sequences. These studies are milestones of development of forecasting science. In the last century forecasting science has been improved and forecasting studies have been carried out in different subjects. On today's complicated and unpredictible environment, in order to increase the efficiency of forecasting studies, people begun to benefit from fuzzy logic approach. In this study, two actual applications in two different subjects which are forecasting the time series of Turkey's foreign trade volume and forecasting the time series of Doğancı Dam's monthly water volume, respectively have been conducted. While creating the fuzzy forecasting models, ARIMA and seasonal ARIMA methodologies have been applied. To find fuzzified forecasts, model parameters have been assumed as triangular fuzzy numbers. By solving the presented linear programming models, spread values have been determined. After determining the width values of parameters, crisp parameters have been transformed into fuzzy parameters. Also by using the fuzzy parameters, upper and lower bounds have been calculated. So, upper and lower bounds can provide the decision makers the best and the worst possible cases. The effects of changes in the time series to upper and lower bounds have been examined.