摘要
通过建立多项式拟合模型找出影响预测结果的异常数据,剔除后建立GM(1,1)模型。对某集团1991年-2003年的伤亡事故统计数据,运用MATLAB工具箱,由图形观测和相对误差分析,提高了模型预测的准确性和适应性,其预测精度大幅提高,预测期望值高于单一的多项式拟合和灰色预测模型。
Abnormal data affecting prediction results is detected by establishing a polynomial fitting model, and GM （ 1, 1 ） model is established after elimination. With respect to the statistic casualties of a group company in the period of 1991 - 2003, we use the MATLAB toolbox to make graphical observation and relative error analysis, thus improving the accuracy and adaptability of model prediction. Its prediction expectation is higher than that of a single polynomial fitting or grey prediction model.
出处
《电气自动化》
2016年第1期12-14,22,共4页
Electrical Automation
基金
甘肃省财政厅专项资金立项资助(甘财教【2013】116号)
