Initial Values for Gauss functions
DESCRIPTION:
Compute initial values for the mean and covariance for the functions
emGauss
and
daGauss
.
USAGE:
Gauss.start.default(data, tol = sqrt(.Machine$double.eps))
Gauss.start.preGauss(object, tol = sqrt(.Machine$double.eps))
REQUIRED ARGUMENTS:
- data
-
a matrix or data frame, normally with missing values.
- object
-
a class
"preGauss"
object.
- tol
-
lower limit on the diagonal entries of the covariance matrix.
VALUE:
a list containing the following components:
- mu
-
column means for the data ignoring missing values.
- sigma
-
diagonal matrix of column variances for the data ignoring missing values.
Variances less than or equal to
tol
are replaced by 1.
REFERENCES:
Schafer, J. L. (1997), Analysis of Incomplete Multivariate Data,
Chapman & Hall, London.
SEE ALSO:
,
,
,
.
EXAMPLES:
Gauss.start(object = cholesterol)
Gauss.start.default(object = cholesterol) # same
cholesterol.pre <- preGauss(cholesterol)
Gauss.start(object = cholesterol.pre)
Gauss.start.preGauss(object = cholesterol.pre) #same