Abstract—In grey forecasting model, the average relative percentage error is an important criterion for assessing the forecasting precision. However, it seldom considers in-sample and out-of-sample tests at the same time. Besides, overestimation or underestimation usually exists when the original data has a rising or falling trend. Thus, this study proposes weighted average nonlinear grey Bernoulli model to solve above two problems. A weighted average moderates the effect of overestimation or underestimation. A new criterion which considers in-sample and out-of-sample tests at the same time indicates that the weighted average nonlinear grey Bernoulli model has the smallest modeling error. Finally, the proposed method is used to forecast Taiwan’s GDP. The results show that Taiwan’s GDP is steadily growing.It may serve as valuable information for policy makers and investors.
Index Terms—Grey forecasting, nonlinear grey bernoulli model, GDP, weighted average.
Pei-Han Hsin is with the Department of International Business, Cheng Shiu University, Kaohsiung City, Taiwan (R.O.C.) (e-mail: phhsin@gmail.com).
[PDF]
Cite: Pei-Han Hsin, "Forecasting Taiwan’s GDP by the Novel Weighted Average Nonlinear Grey Bernoulli Model," International Journal of Modeling and Optimization vol. 5, no. 6, pp. 381-384, 2015.
