MANE 6313

Week 13, Module D

Student Learning Outcome

• Select an appropriate experimental design with one or more factors,
• Select an appropriate model with one or more factors,
• Evaluate statistical analyses of experimental designs,
• Assess the model adequacy of any experimental design, and
• Interpret model results.

Module Learning Outcome

Employ rsm() to design and analyze RSM models.

---

RSM Design in Package rsm()

• rsm() supports two designs:
• Box-Behnken design with 2 to 7 factors
• Center Composite Design (ccd)
• Details and examples of using both designs will be provided in separate modules

---

Model Types

• The model formula for designs in rsm() is different
• FO() specifies first-order model
• TWI() is used to generate two-way interations
• PQ() is used to add pure quadratic terms to model
• SO() creates all terms (FO(),TWI(),PQ()) in a model

---

Coded Variables

• When analyzing RSM models, variables should be coded
• The R chunk for creating coded variables in a model is shown below

library(rsm)
bb3.design <- bbd(3,n0=2,coding=list(x1~(Force-20)/3,x2~(Rate-50)/10,x3~Polish-4))
print(bb3.design)

---

Coded Variables Output

---

Adding a Response Variable

y <- rnorm(14)
bb3.design$y <- y
print(bb3.design)

---

Adding a Response Variable Output

---

Model Fitting

bb3.fitted <- rsm(y~SO(x1,x2,x3),data=bb3.design)
summary(bb3.fitted)

---

Model Fitting Output