线性组合预测的推广  

Generalization of linear combining forecasting

作　　者：高尚[1] 

机构地区：[1]江苏科技大学计算机科学与工程学院,江苏镇江212003

出　　处：《中南大学学报（自然科学版）》2013年第S2期178-182,共5页Journal of Central South University:Science and Technology

基　　金：人工智能四川省重点实验室开放基金资助项目(2009RY001);"青蓝工程"资助项目(2010年)

摘　　要：对传统的线性组合预测模型进行研究,根据权重的约束,对"误差平方之和最小"、"误差绝对值之和最小"和"最大误差绝对值最小"准则的线性组合预测进行推广,补充了若干个线性组合预测模型。以美国加州电力日均价为例,给出9种线性组合预测模型的预测结果。The conventional linear forecasting models were researched.According to the weight constraints,the linear combining forecasting models of 'the least the sum of squares of error','the least the sum of error absolute value' and 'the least the error absolute value' were generalized.The novel several linear combining forecasting models were put forward.Taking the California history daily electricity price as a study case,the prediction results are given by the nine linear forecasting models.

关 键 词：组合预测 误差 电价 

分 类 号：TM711[电气工程—电力系统及自动化] O242.1[理学—计算数学]