Abstract
We consider a linear regression model in continuous time with predetermined stochastic regressors and a local martingale as noise process. The method of estimation of the regression parameters is inspired by the classical least-squares estimate in the discrete time setting. It is shown that under a certain condition limiting the growth of the maximal eigenvalue of the design matrix with respect to the minimal eigenvalue, this “quasi-least-squares” estimate converges with probability one to the true parameter values. The proof uses results from semimartingale theory, in particular the transformation formula and some stability properties of local martingales.

This publication has 7 references indexed in Scilit: