Fuzzy Forecasting Model Based on the Cause-effect Clustering and the Pattern Recognition and Its Application

摘要

We studied the problem of forecasting staff of an organization in the uncertainty environment. In case that the radical change didn' t take place, the factors that influenced the forecasting variable indicated the state of the organization, and every period of history provides a set of the factors. The historic data was clustered by means of fuzzy clustering so as to classify them to many patterns, in which corresponding eigenfunctions were made. eigenvalues were gained by calculating the eigenfunctions with the value of the factors in forecasting period, we could judge the pattern it belonged to according to the principle of max degree of subjection. The triangle fuzzy value in the pattern is used as the forecasting function. In this model, the weights were used to indicate the different influence of the factors. This method can be used in other area.