针对传统GM(1,1)模型建模方法存在偏差,当发展系数的绝对值较大时,模型偏差较大,无法用于中长期预测,甚至不能做短期预测等问题.从优化GM(1,1)背景值出发,通过提高背景值的构造精度,实现对传统GM(1,1)模型的改进,利用MATLAB软件编制高效的计算机程序,实现模型的模拟和预测,并将其应用于我国危险化学品事故起数及死亡人数预测.实例结果表明,改进GM(1,1)模型无论对变化平缓的低增长序列,还是变化急剧的高增长序列,都有较好的适应性,及更好的拟合和预测效果,为提高建模精度提供了新的途径.%There are some problems in GM(1,1) model, such as, model method biased, when the absolute value of development coefficient is larger, the model deviation is larger, it can not be used for mid-long term predictions, and even short-term predictions. The improvement of the traditional GM(1,1) model was implemented by optimizing the GM( 1,1) background value, and improving its structural accuracy, an effective computer program compiled by MATLAB( Matrix Laboratory) software was used to fulfill simulation and prediction of the model, and it was used for predicting the number of accidents and deaths of hazardous chemicals in China. Actual example results showed that the improved GM( 1,1) model had better adaptability, fitting and predicting efficiency for both low growth series with gentle change and high growth series with rapid change, it provides a novel direction to higher modeling accuracy.
展开▼
