Workshop on Statistical Methods for Dynamic
Vancouver, June 4-6 2009
System Models
This presentation will discuss R capabilities for fitting nonlinear differential equations to data. This will include computing both Wald and likelihood confidence intervals and regions for parameter estimates. It will include three general topics: (1) A brief overview of contributed packages for differential equations including ordinary differential equation (ODE), differential algebraic equations (DAE), delay differential equations, stochastic differential equations, and state space models. (2) An analysis of a continuously stirred tank reactor (CSTR) in both R and Matlab. (3) Wald and likelihood methods for confidence intervals, noting that both intrinsic and parameter effects curvature affect Wald intervals, while likelihood intervals are affected only by the intrinsic curvature of the solution manifold.
Authors: Spencer Graves, Giles Hooker and James Ramsay
Fitting Nonlinear Differential Equations to Data in R
