Diffusion processes are used to model a wide range of real-world phenomena.
Consequently, the ability to estimate the diffusion parameters from available
time series data would greatly further our understanding of such phenomena. A
new algorithm is proposed to estimate the parameters for a Fokker-Planck
system. The algorithm is applicable for a wide class of multivariate diffusion
systems. Not only does the method provide reliable parameter estimates in a
highly computationally efficient manner, it can also produce credibility
intervals for these parameters. It is demonstrated that the credibility
intervals have coverage close to their advertised values. Hence, the algorithm
produces parameter estimates with limited systematic biases.