Predicting Beijing's tertiary industry with an improved grey model

Abstract

In the context of the growth slowdown in China, it is important to accurately forecast the future economic trend to guide policy-makers the direction of adjusting their current economic policies. In this paper, we intend to predict Beijing's tertiary industry, whose datasets are small, irregular and non-stationary, leading to a difficulty of building an accurate prediction model. To this end, we present an improved grey model, named PRGM(1,1), which extends the grey prediction model by integrating two techniques, i.e., the particle swarm optimization algorithm for parameter optimization and the exponential preprocessing method for data cleaning. The experimental results show that PRGM(1,1) outperforms other variants of the grey prediction model in predicting Beijing's tertiary industry, and is viable to do reasonable prediction over short and fluctuated economic data sequences. In addition, we employ PRGM(1,1) in the economic prediction of Beijing's tertiary industry in the next five years, and conclude that the growth rate will decelerate. Our prediction result seems to be in line with the economic slowdown in China this year.