An Alternative Estimation Approach for a Multiplicative Latent Process Model

• Main Author: 羅千惠
• Other Authors: 丘政民
• Format: Others
• Language: en_US
• Published: 2001
• Online Access: http://ndltd.ncl.edu.tw/handle/14283415823217318398

碩士 === 國立中正大學 === 數理統計研究所 === 89 === In this study, an alternative estimation approach is proposed for a class of multiplicative latent process models. The main interest is in the marginal model which takes the latent process into account for the consideration of estimation efficiency. We combine the nonparmetric method for variance function estimation and the parametric approach in estimating the correlation structure. These methods are then applied in conjunction with the generalized estimation equations approach iteratively under the framework of generalized linear models. Simulation studies are conducted to examine the performance of the proposed estimation method.