The development and validation of a fuzzy logic method for time -series extrapolation

Jeffrey Stewart Plouffe, University of Rhode Island

Abstract

It has been established that statistically simple extrapolative forecasting methods provide more accurate ex ante forecasts, than do ones that are statistically sophisticated. The problem is that scholars attempting to develop new, more accurate forecasting methods have all but ignored this knowledge on forecast accuracy, (Fildes and Makridakis 1995, Fildes et al, 1998, Makridakis Hibon, 2000. Makridakis and Hibon, (2000), Fildes (2001) and Small and Wong (2002) suggest that what is needed are new statistically simple extrapolative forecasting methods that are robust to the fluctuations that occur in real-life business data. This research discusses the development and validation of the Direct Set Assignment (DSA) extrapolative forecasting method. The DSA method is a new, statistically simple, non-linear, rule based fuzzy logic extrapolative forecasting method that was developed within the Mamdani Development Framework, and has been hypothesized to provide more accurate ex ante forecasts than alternative statistically simple extrapolative methods. The relative forecast accuracy of the DSA method was established in three forecasting competitions that relied on the procedures, standards and a subset of the time series used in the M3 International Forecasting Competition. The DSA method, and the DSA method in combination with Winter's Exponential Smoothing, together provided the highest observed forecast accuracy in seven of the nine subcategories and two of the three categories of time series as well as for all one hundred thirty time series, examined in this study. Also, the DSA method in combination with Winter's Exponential Smoothing provided the highest observed accuracy for the time series in this study in which a statistically significant trend was present. It can be concluded from these findings that the DSA method provides forecasts that are at least as accurate as the alternative extrapolative methods compared in this study. These finding however cannot be generalized beyond this study due to the limitation inherent in the forecasting competition methodology. The next step to advance the DSA method is to develop the means for the a priori identification of the DSA methods fuzzy set parameter. ^

Subject Area

Business Administration, Management

Recommended Citation

Jeffrey Stewart Plouffe, "The development and validation of a fuzzy logic method for time -series extrapolation" (2005). Dissertations and Master's Theses (Campus Access). Paper AAI3186918.
http://digitalcommons.uri.edu/dissertations/AAI3186918

Share

COinS