PubAg

Main content area

A spatial latent class model

Author:
Lee, Jiyon
Source:
Economics letters 2018 v.162 pp. 62-68
ISSN:
0165-1765
Subject:
Monte Carlo method, algorithms, dynamic models, economics
Abstract:
This paper applies the latent class approach to underlying spatial dynamic models. Spatial weights are assumed to be time-varying, although the class membership of each unit is fixed over time. The class-specific parameters and class memberships are estimated with EM algorithm. The performance of the proposed model with a finite sample is examined using Monte Carlo simulation.
Agid:
6101237