Question

The most common way to optimize a regression model in DOE is to add and remove factors terms that have very little significance, then reanalyze the model while checking remaining degrees of freedom and watching the adjusted R^2 and remaining factor p values go up and down. In other words, the most common way to optimize a regression model in DOE is through trial and error. true or false

Answer 1

Correct option:

True

EXPLANATION:

Linear regression models are built through trial and error to balance many competing goals.

The steps to optimize a regression model in DOE areas follows:

1. Examine the variables looking for outliers and making transformations.

2. Construct piecewise scatter plot for each variable. Delete redundant variables.

3. See if new variables need to be brought in.

4. Ft the full Ordinary Least Square model and delete variables with insignificant t - tests.

The above steps are repeated through trial and error procedure.