Log-likelihood computation for Branching Processes with negative binomial noise

[I-Bouros]

• [x] Write down mathematically the log-likelihoods for Poisson and negative binomial noise branching processes

• [x] Create a Pints.LogPDF class for the Poisson Log-likelihood

• [x] Create a Pints.LogPDF class for the negative binomial Log-likelihood with 2 different methods:

  • [x] Overdispersion is a parameter to be learned
  • [x] Overdispersion is a fixed parameter

Use evaluateS1 method from Pints for all these classes! Sliding window length is a fixed parameter on the class which needs to be set!

[I-Bouros]

We split this into two PRs, 1 for Poisson and 1 for negative binomial.

[I-Bouros]

Compare new Poisson methods to previous one in branching processes