Estimation of factor models by realization-based and approximation methods
In this paper we discuss two methods for the estimation of linear dynamic factor models. The first method is behavioural in nature and consists of the least squares approximation of the observed data by means of a linear system. The second method is based on the statistical concept of principal components and uses subspace ideas from approximate realization theory. The two methods are compared by means of simulated data.
|Keywords||Linear system, behaviour, model reduction, principal components, subspace method, system identification|
Scherrer, W., & Heij, C.. (1998). Estimation of factor models by realization-based and approximation methods (No. EI 9831). Retrieved from http://hdl.handle.net/1765/1543