unbinned_extended_maximum_likelihood¶
-
minkit.unbinned_extended_maximum_likelihood(pdf, data, range='full')[source]¶ Definition of the unbinned extended maximum likelihood FCN. In this case, entries in data are assumed to follow a Poissonian distribution. The given
PDFmust be of extended type. The output is two times the logarithm of the likelihood, and is computed as\[\text{FCN} = -2 \times \left[\sum_{i=0}^N \log f(\vec{x}_i;\vec{\theta}) - A(\vec{\theta})\right],\]where \(f(\vec{x}_i;\vec{\theta})\) is the function to minimize (without normalization), \(\vec{x}_i\) are the data points, \(\vec{\theta}\) are the function parameters and \(A(\vec{\theta})\) is the normalization value.