|
- Question:
- What is best transformation of Y so that T(Y) vs. X is
most linear?
- Definition:
- Vertical Axis : Correlation coefficient of T(Y) and X
-   where T(Y) is the Box-Cox transformation
of Y
- Horizontal Axis: Box-Cox power transformation parameter
(lambda)
- Interpretation/Conclusion:
- Best transformation is lambda = 0.6. Note that the
square root transformation (lambda = 0.5) is a reasonable
choice in this case.
- NIST/SEMATECH Handbook Chapter:
- 1.
Exploratory Data Analysis
- 4.
Process Modelling
- Reference:
- Filliben
|
Dataplot Command: BOX-COX LINEARITY PLOT Y X
- Dataplot Code:
- SKIP 25
- READ BERGER1.DAT Y X
- TITLE BOX_COX LINEARITY PLOT
- LABEL CASE ASIS
- Y1LABEL Correlation Coefficient
- X1LABEL Alpha
- BOX-COX LINEARITY PLOT Y X
- Dataplot Reference
Manual Documentation for BOX-COX LINEARITY PLOT
|
Privacy
Policy/Security Notice
Disclaimer |
FOIA
NIST is an agency of the U.S.
Commerce Department.
Date created: 6/5/2001
Last updated: 10/30/2013
Please email comments on this WWW page to
alan.heckert@nist.gov.
|