class TGraphAsymmErrors: public TGraph

 TGraphAsymmErrors default constructor.

 TGraphAsymmErrors copy constructor

 TGraphAsymmErrors assignment operator

 TGraphAsymmErrors normal constructor.

 the arrays are preset to zero

 TGraphAsymmErrors normal constructor.

 if exl,h or eyl,h are null, the corresponding arrays are preset to zero

 TGraphAsymmErrors normal constructor.

 if exl,h or eyl,h are null, the corresponding arrays are preset to zero

 Constructor with six vectors of floats in input
 A grapherrors is built with the X coordinates taken from vx and Y coord from vy
 and the errors from vectors vexl/h and veyl/h.
 The number of points in the graph is the minimum of number of points
 in vx and vy.

 Constructor with six vectors of doubles in input
 A grapherrors is built with the X coordinates taken from vx and Y coord from vy
 and the errors from vectors vexl/h and veyl/h.
 The number of points in the graph is the minimum of number of points
 in vx and vy.

 TGraphAsymmErrors constructor importing its parameters from the TH1 object passed as argument
 the low and high errors are set to the bin error of the histogram.

 Creates a TGraphAsymmErrors by dividing two input TH1 histograms:
 pass/total. (see TGraphAsymmErrors::Divide)

 TGraphAsymmErrors default destructor.

 apply a function to all data points
 y = f(x,y)

 Errors are calculated as eyh = f(x,y+eyh)-f(x,y) and
 eyl = f(x,y)-f(x,y-eyl)

 Special treatment has to be applied for the functions where the
 role of "up" and "down" is reversed.
 function suggested/implemented by Miroslav Helbich <helbich@mail.desy.de>

This function is only kept for backward compatibility.
You should rather use the Divide method.
It calls Divide(pass,total,"cl=0.683 b(1,1) mode") which is equivalent to the
former BayesDivide method.

 Fill this TGraphAsymmErrors by dividing two 1-dimensional histograms pass/total

 This method serves two purposes:

 1) calculating efficiencies:


 The assumption is that the entries in "pass" are a subset of those in
 "total". That is, we create an "efficiency" graph, where each entry is
 between 0 and 1, inclusive.

 If the histograms are not filled with unit weights, the number of effective
 entries is used to normalise the bin contents which might lead to wrong results.
 

 The points are assigned a x value at the center of each histogram bin.
 The y values are  for all options except for the
 bayesian methods where the result depends on the chosen option.

 If the denominator becomes 0 or pass >  total, the corresponding bin is
 skipped.

 2) calculating ratios of two Poisson means (option 'pois'):


 The two histograms are interpreted as independent Poisson processes and the ratio
 
 The histogram 'pass' is interpreted as n_{1} and the total histogram
 is used for n_{2}

 The (asymmetric) uncertainties of the Poisson ratio are linked to the uncertainties
 of efficiency by a parameter transformation:
 

 The x errors span each histogram bin (lowedge ... lowedge+width)
 The y errors depend on the chosen statistic methode which can be determined
 by the options given below. For a detailed description of the used statistic
 calculations please have a look at the corresponding functions!

 Options:
 - v     : verbose mode: prints information about the number of used bins
           and calculated efficiencies with their errors
 - cl=x  : determine the used confidence level (0<x<1) (default is 0.683)
 - cp    : Clopper-Pearson interval (see TEfficiency::ClopperPearson)
 - w     : Wilson interval (see TEfficiency::Wilson)
 - n     : normal approximation propagation (see TEfficiency::Normal)
 - ac    : Agresti-Coull interval (see TEfficiency::AgrestiCoull)
 - fc    : Feldman-Cousins interval (see TEfficiency::FeldmanCousinsInterval)
 - b(a,b): bayesian interval using a prior probability ~Beta(a,b); a,b > 0
           (see TEfficiency::Bayesian)
 - mode  : use mode of posterior for Bayesian interval (default is mean)
 - shortest: use shortest interval (done by default if mode is set)
 - central: use central interval (done by default if mode is NOT set)
 - pois: interpret histograms as poisson ratio instead of efficiency
 - e0    : plot (in Bayesian case) efficiency and interval for bins where total=0
           (default is to skip them)

 Note:
 Unfortunately there is no straightforward approach for determining a confidence
 interval for a given confidence level. The actual coverage probability of the
 confidence interval oscillates significantly according to the total number of
 events and the true efficiency. In order to decrease the impact of this
 oscillation on the actual coverage probability a couple of approximations and
 methodes has been developped. For a detailed discussion, please have a look at
 this statistical paper:

 Compute Range

 Copy and release.

 Copy errors from fE*** to arrays[***]
 or to f*** Copy points.

 Should be called from ctors after fNpoints has been set

 Set zero values for point arrays in the range [begin, end)

 This function is called by GraphFitChisquare.
 It returns the error along X at point i.

 This function is called by GraphFitChisquare.
 It returns the error along Y at point i.

 Get high error on X.

 Get low error on X.

 Get high error on Y.

 Get low error on Y.

 Print graph and errors values.

 Save primitive as a C++ statement(s) on output stream out

 Set ex and ey values for point pointed by the mouse.

 Set ex and ey values for point number i.

 Set EXlow for point i

 Set EXhigh for point i

 Set EYlow for point i

 Set EYhigh for point i

 Stream an object of class TGraphAsymmErrors.

 Swap points.