Filterlng of hidden diffusion processes

From MaRDI portal





The author considers the scalar diffusion equation \[ dx_t= f(t, x_t) dt+ g(t, x_t) dW_t,\quad x_0= 0,\quad 0\leq t\leq 1, \] where \(W_t\) is the Wiener process. The observation process is \(Y_t= x_t I_{x_t< 0}\) (that is, the observation process is invisible when it has nonnegative values, otherwise it is identical to the signal process). Under certain assumptions the filtering equation \(E[f(x_1)\mid x_s I_{x_s< 0}, 0\leq s\leq 1]\) is derived where \(f\) is any bounded measurable function. An example of the technique is applied to the process \(dx_t= dW_t\).











This page was built for publication: Filterlng of hidden diffusion processes

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2747864)