Computing Densities for Stochastic Differential Equations

We give an overview of density tracking by quadrature (DTQ), a method to compute densities for stochastic differential equations. DTQ is motivated by parameter estimation/inference problems, in which it is crucial to compute log likelihoods consisting of density function evaluations. After reviewing DTQ's derivation, we outline DTQ's theoretical and computational properties. We give multiple code samples to explain how to use DTQ in R, especially with our open-source Rdtq package.