Estimation of Diffusion Parameters by a Nonparametric Drift Function Model
The paper presents a method to estimate diffusion parameters without specifying drift functions of one dimensional stochastic differential equations. We study finite sample properties of the estimator by numerical experiments at several observation time intervals with total time interval fixed. The results show the estimator is getting efficient as observation time interval becomes smaller. By comparing with the quadratic variation method which is proven to have consistency, the proposed method produces almost the same finite sample properties as that.
KeywordsObservation Time Interval Discretized Process Drift Function Discrete Time Series Total Time Interval
Unable to display preview. Download preview PDF.
- 11.Tjøsthaim, D., Auestad, B.H.: Nonparametric identification of nonlinear time series: projections. J. Am. Statist. Assoc. 89, 258–266 (1994)Google Scholar