This is a highly fractionated two-level factorial design employed as a screening design in an off-line welding experiment performed by the National Railway Corporation of Japan. There were 16 runs and 9 experimental factors. The response variable is the observed tensile strength of the weld, one of several quality characteristics measured. All other variables are at plus and minus levels.
data(welding)
A data frame containing the following variables.
All the explanatory variables are numeric with two levels, -1
and 1
.
Variable | Description | |
Rods | Kind of welding rods | |
Drying | Period of drying | |
Material | Welded material | |
Thickness | Thickness | |
Angle | Angle | |
Opening | Opening | |
Current | Current | |
Method | Welding method | |
Preheating | Preheating | |
Strength | Tensile strength of the weld in kg/mm. The response variable. |
Smyth, G. K., Huele, F., and Verbyla, A. P. (2001). Exact and approximate REML for heteroscedastic regression. Statistical Modelling 1, 161-175.
Smyth, G. K. (2002). An efficient algorithm for REML in heteroscedastic regression. Journal of Computational and Graphical Statistics 11, 1-12.