Parameter estimation by a Markov chain Monte Carlo technique for the Candy model
Xavier Descombes, M.N.M. van Lieshout, Radu Stefan Stoica, Josiane B. Zerubia · 2002
This paper presents a parameter estimation method for the Candy model based on Monte Carlo approximation of the likelihood function. In order to produce such an approximation a Metropolis-Hastings style algorithm for simulating the Candy model is introduced.