首页> 外国专利> GENERATING METHOD OF TWO CLASS CLASSIFICATION PREDICTION MODEL, PROGRAM FOR GENERATING CLASSIFICATION PREDICTION MODEL AND GENERATING DEVICE OF TWO CLASS CLASSIFICATION PREDICTION MODEL

GENERATING METHOD OF TWO CLASS CLASSIFICATION PREDICTION MODEL, PROGRAM FOR GENERATING CLASSIFICATION PREDICTION MODEL AND GENERATING DEVICE OF TWO CLASS CLASSIFICATION PREDICTION MODEL

机译:两类分类预测模型的生成方法,两类分类预测模型的生成程序和生成装置

摘要

A kind of two-stage classification/prediction generates in model, by executing simple two classes classification and classification rates substantially close to 100%. What the two-stage classification/prediction model generated) explanatory variable that is generated on the basis of predetermined class is other into two for the training sample set of classification of acquisition discriminant function, to the training sample set for including in each sample, b) each training sample of computational discrimination score, utilize gained discriminant function, c) it determines, based on discriminant scores calculated, regardless of whether being the training sample correctly classified, d) determine that wrong classification samples are derived from the training sample set based on pole maximum and minimum discriminant scores in wrong classification samples, e) new training sample set is constructed, wrong classification samples area is included in by extracting training sample, f) repeat) to e) for new training sample set.
机译:通过执行简单的两类分类和分类率基本上接近100%,可以在模型中生成一种两阶段分类/预测。基于获取的判别函数的分类的训练样本集,针对每个样本中包含的训练样本集,将基于预定类别而生成的两阶段分类预测模型的解释变量另分为两个。 )计算辨别力分数的每个训练样本,利用获得的判别函数,c)根据计算出的判别分数确定,无论训练样本是否正确分类,d)确定基于训练样本集得出的错误分类样本在错误分类样本的极点最大和最小判别分数上,e)构建新的训练样本集,通过提取训练样本将错误的分类样本区域包括在内,f)对新训练样本集重复e)到e)。

著录项

  • 公开/公告号KR20100082377A

    专利类型

  • 公开/公告日2010-07-16

    原文格式PDF

  • 申请/专利权人 FUJITSU LIMITED;

    申请/专利号KR20107013416

  • 发明设计人 YUTA KOHTAROU;

    申请日2007-12-18

  • 分类号G06N5/04;G06F15/18;G06N3/00;

  • 国家 KR

  • 入库时间 2022-08-21 18:32:13

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