Statistics for continuous-time stochastic processes in a nutshell
A continuous-time stochastic process can be thought of as a collection of random variables { X(t), t > 0 }. Such processes are widely used for modelling phenomena that evolve over time. For instance, one can think of modelling a population size, where at certain random times, an individual might die or a new individual is born, but there are many more examples, such as solutions to SDEs or chemical reaction processes. As is usually the case in mathematical models, there are almost always some underlying parameters governing the dynamics of the process, which are of interest. In the talk, I’ll introduce some notions from the field of stochastic processes and then I’ll discuss the statistical problem of estimating such parameters based on a discrete set of observations { X(t_k), k = 1, … n }. If time permits, I’ll demonstrate a method proposed in our latest article in which we propose a simulation method for the so-called bridge process between the observations.