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