how do I identify the significant factors using Plackett Burman Design in Matlab?

The system is made up of 384 factors. For the Plackett Burman Design I used the code hadamard() to generate the matrix. The matrix have 384 runs. The attachment gives an example how the matrix roughly looks like. The values of activity are given. I would like to use the code stepwise() to count the p-value. Is this method applicable in my case? How does it work with the set of values of activity? or any suggestion how to do the design?

Thanks in advance.

[John BG]

please correct if wrong,

1.- you want extrapolate the probability density function out of the the runs.

2.- you are assuming Weibull pdf

3.- the p-value you mention is the gamma g, shape parameter, used in some literature to define the standard Weibull (not the general) pdf as follows:

f(x)=g*x^(x-1)*exp(-x^g)

x>=0 g>0

4.- you are not considering the general Weibull probability distribution:

f(x)=g/a*((x-u)/a)^(g-1)*exp(-((x-u)/a)^g)

g: shape parameter

u: location parameter

a: scale parameter

awaiting answer

John

[e-mail]