...
首页> 外文期刊>Quality Control and Applied Statistics >Using individual factor information in fractional factorial designs
【24h】

Using individual factor information in fractional factorial designs

机译:Using individual factor information in fractional factorial designs

获取原文
获取原文并翻译 | 示例
           

摘要

To investigate factorial effects, two-level factorial designs are widely used with a focus on having designs effective and efficient designs. However, not much attention is paid to the way columns are assigned to factors once the design is selected. Li et al. (Ref. 1) proposed the individual word length pattern (iWLP) to measure aliasing severity of columns to factors on selection of the design. In regular designs, the effects are either orthogonal or fully aliased. In nonregular designs, many factors will be partially aliased. This can cause bias due to omission of interactions, and decrease in precision due to correlated contrast for aliased terms. This study uses individual generalized word length pattern (iGWLP) for as a measure for the nonregular designs. The iGWLP criterion can be used to reveal additional information about factors involved in aliasing. If prior information is available about any of the factors, then it could be suitably incorporated in the design so that maximum details can be gathered. Another contribution of this study is providing a theoretical justification for iGWLP and iWLP. This is related to focusing on partially aliased effects where the G2aberration criterion sequentially minimizes contamination from negligible interactions (Ref. 2.) Jones and Nachtsheim (Ref. 3) targeted minimizing the bias caused by missing two-factor interactions, estimating intercept and main effects. This study measures the aliasing interactions using a new criterion. The third contribution of this study is use of both iWLP and IGWLP for design selections.

著录项

获取原文

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