Page 252 - Six Sigma Demystified

232 Six SigMa DemystifieD

Excel

Use Data\Data Analysis\Random Number Generation.
Set Distribution = Normal. Transform normally distributed data to lognormal
using the natural log function. For example, if cell A2 contains normally distrib-
uted data, LN(A2) contains lognormally distributed data.

Weibull Distribution

Used for measurement (continuous) data that theoretically are without bound
in only the positive direction and bounded at zero (or a positive value above
zero), the Weibull distribution is often applied to reliability (time to failure)
and financial analyses.

Weibull Distributions

Minitab

Use Calc\Random Data\Weibull to generate random numbers. Specify shape,
scale, and threshold (which equals zero for a two-parameter distribution).
Use Stat\Quality Tools\Individual Distribution Identification to test whether the
sample data meet the Weibull distribution. Use goodness-of-fit tests (described

below) to determine if an assumed distribution provides a reasonable
approximation.

Excel

Use Data\Data Analysis\Random Number Generation.
Set Distribution = Uniform. Transform the uniformly distributed data to
Weibull using the natural log function and shape and scale parameters. For
example, if cell A2 contains uniformly distributed data, =15*(–LN(A2))^(1/3)
contains Weibull distributed data with a shape parameter of 15 and a scale
parameter of 3.