mpcv (version 1.1)

mpcv: Multivariate process capability vector

Description

Performs the multivariate process capability analysis using three component multivariate process capability vector (mpcv).

Usage

mpcv(x, indepvar = 1, LSL, USL, Target, alpha = 0.0027, distance, n.integr = 100,  coef.up, coef.lo)

Arguments

x

a numeric matrix containing the data (quality characteristics).

indepvar

a number or a name of the independent variable needed for building one-sided models.

LSL

a vector of lower specification limits defined for each variable.

USL

a vector of upper specification limits defined for each variable.

Target

a vector of target of the process defined for each variable.

alpha

the proportion of nonconforming products.

distance

the distance measure to be used for removing the nonconforming elements. This must be one of "mahalanobis" (default), "euclidean", "maximum", "manhattan", "canberra", "binary" or "minkowski". Any unambiguous substring can be given.

n.integr

a number of integration intervals

coef.up

a vector of minimal values of leading coefficients for "upper" one-sided models. Value given for indepvar is omitted (could be NA).

coef.lo

a vector of minimal values of leading coefficients for "lower" one-sided models. Value for indepvar is omitted (could be NA).

Value

An mpcv object. See mpcv.object for details.

Details

If the parameter Target is not specified, then Target <- LSL + (USL - LSL)/2.

Using the applied methodology, the shape of a process region is mainly defined by the leading coefficients of the models. To obtain a certain shape of a process region (e.g. similar to the previous one) there is possible to specify minimal values of the leading coefficients coef.lo and coef.up of the models. By default all the minimal values of the coefficients equal zero.

Except the "mahalanobis" distance, the available distance measures are listed in dist.

References

Ciupke K. (2014) Multivariate Process Capability Vector Based on One-Sided Model, Quality and Reliability Engineering International, John Wiley & Sons.

Examples

data(industrial)   
x <- industrial$x
LSL <- industrial$LSL
USL <- industrial$USL
Target<- industrial$Target
res.ind <- mpcv(x, LSL=LSL, USL=USL, Target=Target, alpha=0.025)

data(automotive)
x <- automotive$x
LSL <- automotive$LSL
USL <- automotive$USL
Target<- automotive$Target
res.aut <- mpcv(x, indepvar="T",  LSL=LSL, USL=USL, Target=Target)

data(sleeves)
x <- sleeves$x
LSL <- sleeves$LSL
USL <- sleeves$USL
Target<- sleeves$Target
res.sle <- mpcv(x, indepvar=3, LSL=LSL, USL=USL, Target=Target, alpha=.02)