A prediction method for irregular demand processes

Irregular demand forecasting method based on radial basis function network (RBFN) is presented in this paper. Firstly, some researches in the literatures are discussed, and then a modeling prediction method based on RBFN is proposed for irregular demand time series. According to the irregularity of time series, the method can successively adjust model structure and parameters by error, so that can effectively simulate rules of system; Meanwhile, the orthogonality and forward regression can realize the recursive computation, greatly reduce the amount of computation and more satisfy practical desires. Finally the effectiveness of the method is illustrated by the simulation examples.