Affiliation of Author(s):经济与管理学院
Journal:COMMUNICATIONS IN NONLINEAR SCIENCE AND NUMERICAL SIMULATION
Key Words:Grey-Markov model Background value Whitenization weighted function Grey prediciton Markov chain
Abstract:Grey-Markov forecasting model is a combination of grey prediction model and Markov chain which show obvious optimization effects for data sequences with characteristics of non-stationary and volatility. However, the state division process in traditional Grey-Markov forecasting model is mostly based on subjective real numbers that immediately affects the accuracy of forecasting values. To seek the solution, this paper introduces the central-point triangular whitenization weight function in state division to calculate possibilities of research values in each state which reflect preference degrees in different states in an objective way. On the other hand, background value optimization is applied in the traditional grey model to generate better fitting data. By this means, the improved Grey-Markov forecasting model is built. Finally, taking the grain production in Henan Province as an example, it verifies this model's validity by comparing with GM(1,1) based on back-ground value optimization and the traditional Grey-Markov forecasting model. (C) 2017 Elsevier B.V. All rights reserved.
ISSN No.:1007-5704
Translation or Not:no
Date of Publication:2018-01-01
Co-author:叶璟,yejing
Correspondence Author:叶璟,dyg,yejing
Date of Publication:2018-01-01