Error modification of grey models using principle of concatenation

In literature a number of different methods are proposed to improve the prediction accuracy of grey models. However, most of them are computationally expensive, and this may prohibit their extensive use. This paper describes a much simpler scheme, based on the principle of concatenation, in which unit step predictions are concatenated by replacing the missing outputs by their previously predicted values. Despite its extreme simplicity, it is shown that the predicted values thus derived results in a better performance than the methods proposed in the literature. Simulation studies show the effectiveness of the proposed algorithm when applied to a chaotic function prediction.