parabola.py 文件源码

def bayesianUpperLimit(self, alpha, steps=1.e5, plot=False):
        """
        Compute one-sided upper limit using Bayesian Method of Helene.
        Several methods of increasing numerical stability have been implemented.
        """
        x_dense, y_dense = self.densify()
        y_dense -= numpy.max(y_dense) # Numeric stability
        f = scipy.interpolate.interp1d(x_dense, y_dense, kind='linear')
        x = numpy.linspace(0., numpy.max(x_dense), steps)
        pdf = numpy.exp(f(x) / 2.)
        cut = (pdf / numpy.max(pdf)) > 1.e-10
        x = x[cut]
        pdf = pdf[cut]
        #pdf /= pdf[0]
        #forbidden = numpy.nonzero(pdf < 1.e-10)[0]
        #if len(forbidden) > 0:
        #    index = forbidden[0] # Numeric stability
        #    x = x[0: index]
        #    pdf = pdf[0: index]
        cdf = numpy.cumsum(pdf)
        cdf /= cdf[-1]
        cdf_reflect = scipy.interpolate.interp1d(cdf, x)

        #if plot:            
        #    pylab.figure()
        #    pylab.plot(x, f(x))
        #    pylab.scatter(self.x, self.y, c='red')
        #    
        #    pylab.figure()
        #    pylab.plot(x, pdf)
        #    
        #    pylab.figure()
        #    pylab.plot(cdf, x)

        return cdf_reflect(alpha)