摘要
通过分析巴伦诺尔一矿煤质得到了17组煤质分析数据,包括Mad,Ad,Hdaf,Qgr,d。利用多元回归分析的原理,建立煤的发热量关于煤中水分和灰分含量的多元回归方程,并通过R检验、F检验、t检验,证明了回归方程的作用显著,即具有实用价值。但是线性回归分析仅能对煤的发热量进行估算,并不能精确预测。因此,采用了支持向量机(SVM)算法对多元线性回归的初步预测结果进行小范围修正,修正结果显示总体预测精度明显提高,这2种方法的结合,效果优于常用方法。
Get seventeen groups coal analysis data of Balunnuoer NO. 1 coal mine,involving Mad ,Ad, Hdaf and Qgr,d" Deduce the multiple regression equation about the calorific value of coal and Mad, Ad based on the principle of multiple regression analysis. The R test, F test, t test show that the regression equation has practial value. However, the results are not accurate. So the support vector machine (SVM) is adopted to revise the preliminary forecasting results on a small scale. The results show that, through revising, the overall forecasting accuracy is greatly improved and it's more efficient than conventional methods.
出处
《洁净煤技术》
CAS
2012年第3期67-70,共4页
Clean Coal Technology
关键词
煤发热量
多元线性回归
支持向量机
预测模型
calorific value of coal
multiple linear regression
support vector machine
prediction model