semPLS (version 1.0-6)

semplsGOF: Quality Indices and Goodness of fit measures for pls path models

Description

A collection of method to validate the goodness of the model. Since there is no well identified global optimization criterion each part of the model needs to be validated.

Usage

rSquared(object, ...)
## S3 method for class 'sempls':
rSquared(object, na.rm=FALSE, \dots)
## S3 method for class 'rSquared':
print(x, na.print=".", digits=2, \ldots)
qSquared(object, ...)
## S3 method for class 'sempls':
qSquared(object, d=NULL, impfun, dlines=TRUE,
         total=FALSE, \dots)
## S3 method for class 'qSquared':
print(x, na.print=".", digits=2, \ldots)
dgrho(object, ...)
## S3 method for class 'sempls':
dgrho(object, \ldots)
## S3 method for class 'dgrho':
print(x, na.print=".", digits=2, \ldots)
communality(object, ...)
## S3 method for class 'sempls':
communality(object, \ldots)
## S3 method for class 'communality':
print(x, na.print=".", digits=2, \ldots)
redundancy(object, ...)
## S3 method for class 'sempls':
redundancy(object, \ldots)
## S3 method for class 'redundancy':
print(x, na.print=".", digits=2, \ldots)
rSquared2(object, ...)
## S3 method for class 'sempls':
rSquared2(object, na.rm=FALSE, \dots)
## S3 method for class 'rSquared2':
print(x, na.print=".", digits=2, \ldots)
gof(object, ...)
## S3 method for class 'sempls':
gof(object, \ldots)
## S3 method for class 'gof':
print(x, na.print=".", digits=2, \ldots)

Arguments

object

An object of class sempls.

d

A numeric value for the omission distance. Thus choosing $d=N-1$, where $N$ is the number of complete observations, is leaving one out cross validation. This is done when d takes its default value NULL.

impfun

An user specified function to impute missing values.

dlines

If TRUE the same observations are deleted for a whole block of MVs, else each dth , counting from top left to bottom right, observation is deleted.

total

If total=TRUE total effects are used instead of path coefficients to calculate the predictions.

na.rm

If na.rm=TRUE observation with missing values are discarded before analysis.

x

An object of the according class.

na.print

A character substituting values not to be printed.

digits

minimal number of _significant_ digits, see print.default.

...

Arguments to be passed down.

Value

• Most GOF methods return a column vector with the names of the variables as rows and the respective measure as column.

References

Esposito Vinzi V., Trinchera L., Amato S. (2010). PLS Path Modeling: From Foundations to Recent Developments and Open Issues for Model Assessment and Improvement. In Esposito Vinzi V., Chin W.W., Henseler J., Wang H.F. (eds.), Handbook of Partial Least Squares: Concepts, Methods and Applications in Marketing and Related Fields, chapter 2. Springer-Verlag Berlin Heidelberg.

See Also

sempls, plsLoadings

Examples

data(ECSImobi)
ecsi <- sempls(model=ECSImobi, data=mobi, E="C")

### R-squared
rSquared(ecsi)

### Q-squared with omission distance d=4
qSquared(ecsi, d=4)

### Dillon-Goldstein's rho (aka composite reliability)
dgrho(ecsi)

### Communalities
communality(ecsi)

### Redundancy
redundancy(ecsi)

### R-squared (normal + corrected)
rSquared2(ecsi)

### Goodness of fit
gof(ecsi)

### check for discriminant validity using loadings
l <-plsLoadings(ecsi)
print(l, type="discriminant", cutoff=0.5, reldiff=0.2)