Time Series Forecasting with Stochastic Markov Models based on Fuzzy Set and Grey Theory

摘要

The stochastic Markov model is combined with fuzzy set concept and grey system for improving forecasting performance. The data for model test is obtained from ACI including Hong Kong, Beijing, Taoyuan, Incheon and Narita international airport. The empirical results show that fuzzy Markov model has better predictive performance with the data with trend and intercept. For the data with random walk, grey Markov model performs better. The paper also examines the effects of transition state and length of interval on the forecasting performance with the result.