NMixMCMCinity function

Initial values of censored observations for the NMixMCMC function

This is a help function for NMixMCMC function. If inity is not given, it calculates reasonable initial values for censored observations. If inity argument is given then it is checked for consistency and formatted on output.

THIS FUNCTION IS NOT TO BE CALLED BY ORDINARY USERS.

NMixMCMCinity(y0, y1, censor, sd.init,
   are.Censored, are.Right, are.Exact, are.Left, are.Interval,
   p, n, inity, random=FALSE)

Arguments

• y0: see output from NMixMCMCdata
• y1: see output from NMixMCMCdata
• censor: see output from NMixMCMCdata
• sd.init: a vector of length $p$ with initial values for overall standard deviations in each margin
• are.Censored: see output from NMixMCMCdata
• are.Right: see output from NMixMCMCdata
• are.Exact: see output from NMixMCMCdata
• are.Left: see output from NMixMCMCdata
• are.Interval: see output from NMixMCMCdata
• p: dimension of the data
• n: number of observations
• inity: a vector (if $p=1$) or a $n x p$ matrix (if $p >= 1$) of initial values of censored observations to be checked for consistency. If not given then reasonable initials are generated.
• random: logical value. If TRUE then some randomness is used when generating initial values.

Returns

A $n x p$ matrix with reasonable initial values for censored observations.

See Also

NMixMCMC.

Author(s)

Arnošt Komárek arnost.komarek@mff.cuni.cz