class RooStats::MCMCCalculator: public RooStats::IntervalCalculator, public TNamed

 set the model

 Main interface to get a RooStats::ConfInterval.

 default constructor

 Get the size of the test (eg. rate of Type I error)

 Get the Confidence level for the test

 Set the DataSet if not already there

 Set the Pdf if not already there

 Set the Prior Pdf if not already there

 specify the parameters of interest in the interval

 specify the parameters to store in the Markov chain
 By default all the parameters are stored

 specify the nuisance parameters (eg. the rest of the parameters)

 set the conditional observables which will be used when creating the NLL
 so the pdf's will not be normalized on the conditional observables when computing the NLL

 set the size of the test (rate of Type I error) ( Eg. 0.05 for a 95% Confidence Interval)

 set the confidence level for the interval (eg. 0.95 for a 95% Confidence Interval)

 set the proposal function for suggesting new points for the MCMC

 set the number of iterations to run the metropolis algorithm

 set the number of steps in the chain to discard as burn-in,
 starting from the first

 set the number of bins to create for each axis when constructing the interval

 set which variables to put on each axis

 set whether to use kernel estimation to determine the interval

 set whether to use sparse histogram (if using histogram at all)

 set what type of interval to have the MCMCInterval represent

 Set the desired level of confidence-level accuracy  for Keys interval
 determination.

 When determining the cutoff PDF height that gives the
 desired confidence level (C_d), the algorithm will consider acceptable
 any found confidence level c such that Abs(c - C_d) < epsilon.

 Any value of this "epsilon" > 0 is considered acceptable, though it is
 advisable to not use a value too small, because the integration of the
 Keys PDF sometimes does not have extremely high accuracy.

 When the shortest interval using Keys PDF could not be found to have
 the desired confidence level +/- the accuracy (see
 SetKeysConfidenceAccuracy()), the interval determination algorithm
 will have to terminate with an unsatisfactory confidence level when
 the bottom and top of the cutoff search range are very close to being
 equal.  This scenario comes into play when there seems to be an error
 in the accuracy of the Keys PDF integration, so the search range
 continues to shrink without converging to a cutoff value that will
 give an acceptable confidence level.  To choose how small to allow the
 search range to be before terminating, set the fraction delta such
 that the search will terminate when topCutoff (a) and bottomCutoff (b)
 satisfy this condition:

 TMath::Abs(a - b) < TMath::Abs(delta * (a + b)/2)