A1207
Title: Semiparametric estimation for transformation models integrating external aggregate data under population heterogeneity
Authors: Chang-Yu Tsai - National University of Kaohsiung (Taiwan) [presenting]
Abstract: Aggregated information is introduced from external sources to improve estimation efficiency. An empirical likelihood-based framework is proposed for semiparametric transformation models that incorporate t-year subgroup survival probabilities from external data. The proposed estimation procedure synthesizes an additional likelihood component to account for uncertainty in the external information and employs a density ratio model to address population heterogeneity. Both the large-sample properties and simulation results are presented, which show that the proposed estimator is more efficient than conventional approaches. Finally, the practical utility of the method is demonstrated through an application to a pancreatic cancer dataset.
