mgrwc: Gaussian mixture of regression models with contamination
In ekstroem/mommix: Moment-Based Estimation of Regression Mixtures

Description Usage Arguments Details Value Author(s) Examples

Fast computation of simple regression slopes for each predictor represented by a column in a matrix

1 2	mgrwc(y, X, weight, maxit = 400L, tol = 1e-07, alpha = 0.3, mu = NA_real_, mufixed = FALSE)

`y`	A vector of outcomes.
`X`	A design matrix of regressor variables. Must have the same number of rows as the length of y.
`weight`	A vector of weights for each observation that are used in the estimation.
`maxit`	the maximum number of iterations to use for the EM algorithm
`tol`	the tolerance used for determining convergence
`alpha`	the starting probability for an observation originating from the contamination distribution. Must be strictly between 0 and 1
`mu`	the starting value for the mean of the contamination distribution. If set to NA (the default) then the mean of the y vector is used.
`mufixed`	Should the value ...

Missing values (NA, Inf, NaN) are completely disregarded and pairwise complete cases are used for the analysis.

A list with the variables: N, K, coefficients, mu, sigma1, sigma2, alpha, iterationsused, and groupprob which contains information on the nuber of observations, mixture components, coefficients for ...

Claus Ekstrom <ekstrom@sund.ku.dk>

1
2
3

x <- rnorm(1000)
y <- rnorm(1000, mean=c(x[1:700], rep(0, 300)), sd=rep(c(1,2), times=c(700,300)))
mgrwc(y, cbind(rep(1, 1000), x), weight=rep(1, 1000))