Grey number prediction using the grey modification model with progression technique

Abstract

Although the grey model has been employed in various fields and demonstrated promising results, most applications focus on precise number predictions. However, owing to the increasing complexity of real-world problems, the grey number predictions will be more flexible and practical for grey model to describe the uncertain future tendency. For the purpose of establishing a grey model with grey number, this paper proposes a progression technique, which adopts different length series divided from the simulation data to produce a grey number prediction. Besides, this paper also modifies the algorithm of grey model, including altering the calculation of background value with an integration term and replacing the initial value of grey differential equation to the latest point, to enhance its accuracy. Two illustrative examples of numerical series and stock market are adopted for demonstrations. Results show that the proposed model can both catch the future tendency and reduce the loss of erroneous judgments.