Using improved grey forecasting models to forecast the output of opto-electronics industry

Numerous forecasting models have been developed. Each has its own conditions of application. However, it has always been an important research objective to improve prediction accuracy with a small amount of data. In recent years, the grey forecasting model has achieved good prediction accuracy with limited data and has been widely used in various research fields. However, the grey forecasting models still have some potential problems that need to be improved. Therefore, this study proposed an improved transformed grey model based on a genetic algorithm (ITGM(1,1)), and used the output of the opto-electronics industry in Taiwan from 1990 to 2008 as an example for verification. Three grey forecasting models, GM(1,1), rolling GM(1,1), and the transformed GM(1,1), were chosen for the purpose of comparison with ITGM(1,1) by mean absolute percent error and root mean square percent error. The results show that ITGM(1,1) is more accurate than the other three models in both in-sample and out-of-sample forecasting performance, and can greatly improve the accuracy of short-term forecasts.

► We chose Taiwan opto-electronics industry as our research object. ► This study proposed an improved transformed grey model based on a genetic algorithm. ► The ITGM(1,1) is an adequate forecasting tool for small data sets. ► The proposed ITGM(1,1) model is very useful for forecasting the output of the opto-electronics industry in Taiwan.