Forecasting of Egypt Wheat Imports Using Multivariate Fuzzy Time Series Model Based on Fuzzy Clustering

This paper presents Multivariate-Factors fuzzy time series model for improving forecasting accuracy. The proposed model is based on fuzzy clustering and it employs eight main procedures to build the multivariate-factors model. The model is evaluated by studying the Egypt Wheat imports as a forecasting problem. Forecasting Egypt wheat imports depend on three factors: population size, wheat area, and wheat production. This forecasting problem is considered to be a good benchmark for comparing different forecasting techniques since it exhibits highly nonlinearities over a long period of time and it provides important economical indicators needed for national future planning. Experimental results show that the proposed model provides higher forecasting accuracy than ARIMA model, Regression model and neural network model. Therefore, the proposed model can lead to satisfactory high performance for fuzzy time series.