Calculate Signal-to-Noise Ratios for a Robust Design

DESCRIPTION:

Computes summary statistics and signal-to-noise rations over noise factors of responses from a robust design.

USAGE:

robust.sn(robust.df, snratio="target", response) 

REQUIRED ARGUMENTS:

robust.df
object of class "robust.design".

OPTIONAL ARGUMENTS:

snratio
character string specifying the signal-to-noise ratio to compute. snratio can take the values: "small" denotes the smaller the better, "large" denotes the larger the better, and "target", the default, denotes the closer to target the better (nominal-the-best).
response
names of the response variables in the data frame robust.df for computing the signal to noise ratios. The default is all columns not "noise" or "control" factors in robust.df.

VALUE:

Returns an object of class robust.sn which inherits from data.frame. It contains the design points of the control array used to generate the robust.df, and the following quantities for each response in the data frame.
mean
mean response.
sd
standard deviation of the response.
meanl
mean of the log of the response.
lsdl
log standard deviation of the log of the response and the signal to noise ratio. RC target, small or large the signal-to-noise ratio specified.

The computed quantities are named in the returned data frame by appending the response names to each of the above names using a . (dot).

DETAILS:

The summary statistics computed by default are mean ( mean), standard deviation ( sd), mean(log(y)) ( meanl), log(standard deviation(log(y))) ( lsdl). These can then be used as responses to study behavior of the mean and standard deviation, via fac.aov. In addition, one of Taguchis signal-to-noise ratio statistics is computed.

If "target" is specified, then the nominal-is-best ratio is computed: If "smallest", the smaller-is-better ratio is computed: If "largest" is specified, the larger-is-better ratio is computed:

REFERENCES:

Taguchi, G. (1986). Introduction to Quality Engineering: Designing Quality Into Products and Processes, Tokyo: Asian Productivity Organization.

Phadke, M. (1989). Quality Engineering using Robust Design, AT&T Bell Laboratories, Prentice Hall, New Jersey.

SEE ALSO:

, , , .

EXAMPLES:

mold.sn <- robust.sn(mold.df)