MCMCFilter generates a simulated samples for the given network using Markov Chain Monte Carlo method.

NOTE: This implementation is only for test and evaluation purpose only.

Arguments

samples=n
n=n: The number of samples to sample. (default: n = 1000)
burnin=n
b=n: The number of burn-in iterations. (default: n = 1000)
sd=v: Standard deviation of the normal distribution for generating random numbers. A smaller value produces the lower acceptance ratio. Some literatures recommend to tune this parameter for acceptance ratio around 40%. Basically, acceptance ratio of 40% - 60% should be fine for most cases. A bigger model requires the lower acceptance ratio to generate better samplings. (default: v = 0.1)
output=file
out=file
o=file: Output file name. The file is a tab seperated text file where each line represents a sample. The first line is the header representing names of nodes (variables).
input=file
in=file
i=file: Input file name. The input file specifies the fixed (observed) values for variables. This is optional and users do not need to specify all of the values of the variables in the file. If the values are not set for variables, these variables are assumed to not have observed values. Use "NA" to represent unobserved variables.
v=n: Verbose level. (default: n = 0)
help: Shows help and quits.