首页> 外国专利> Compilation method of the forecasting model due to multiple regression analysis, the compilation device, the compilation program

Compilation method of the forecasting model due to multiple regression analysis, the compilation device, the compilation program

机译:基于多元回归分析的预测模型的编制方法,编制装置,编制程序

摘要

The forecasting model of the intended variable where the estimate precision which is based on multiple regression analysis is high is formed with the computer. Until a) survey value of intended variable it prepares early sample set with the known sample, b) multiple regression analyzes this set, acquires the calculated value of intended variable, multiple regression analyzing the subsample set which removes the sample where difference of c) survey value and calculated value is below 1st value, it calculates coefficient of determination, d) coefficient of determination exceeds 2nd value, modifying 1st value, step c) it repeats, e) step d) makes the subsample set when ending 1st subsample set,Doing 2 class classifications remainder as a 2nd subsample set, it calculates discriminant function, f) step d) multiple regression in analyzing when ending system and step e), it sets the discriminant function in g) sets 2nd subsample set to early sample set, step b) from step f) acquires the plural forecasting models over again as the forecasting model.
机译:利用计算机形成基于多元回归分析的估计精度较高的预期变量的预测模型。直到a)预期变量的调查值,它准备与已知样本的早期样本集,b)多元回归分析该样本集,获取预期变量的计算值,多元回归分析子样本集,然后移除样本,c)调查的差异值和计算值低于第一个值,则计算确定系数,d)确定系数超过第二个值,修改第一个值,重复步骤c),e)步骤d)在结束第一个子样本集时形成子样本集, 2类分类余数作为第二个子样本集,它计算判别函数,f)步骤d)分析结束系统和步骤e)时的多元回归,它在g)集合中将判别函数设置为第二个子样本集,作为早期样本集,步骤b)从步骤f)重新获得多个预测模型作为预测模型。

著录项

  • 公开/公告号JPWO2008126209A1

    专利类型

  • 公开/公告日2010-07-22

    原文格式PDF

  • 申请/专利权人 富士通株式会社;

    申请/专利号JP20090508760

  • 发明设计人 湯田 浩太郎;

    申请日2007-03-27

  • 分类号G06N5/04;G06F19/00;

  • 国家 JP

  • 入库时间 2022-08-21 18:57:51

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号