Date of Award
Summer 2022
Document Type
Open Access Thesis
Department
Statistics
First Advisor
Jiajia Zhang
Abstract
Mixture cure model is a useful method of survival analysis for population including cured proportion and uncured proportion. The R package SMCURE applies EM algorithm to estimate the coefficients of covariates in the mixture cure model. Although an offset term is specified in the SMCURE statement, the offset term is not appropriately handled in the algorithm. This thesis aims to adjust the EM algorithm for the proportional hazards mixture cure model in the SMCURE package. In addition, the offset term can be specified separately in the incidence part or the latency part. The numerical experiments include simulation study and real data application on the bone marrow transplantation data, and the results indicate that the modified EM algorithm for the proportional hazards mixture cure model with offset term generates smaller bias and variance estimation, compared to the proportional hazards mixture cure model without considering offset terms.
Rights
© 2022, Jiaying Yi
Recommended Citation
Yi, J.(2022). Modified EM Algorithm in SMCURE Package Based on Proportional Hazards Mixture Cure Model With Offset Terms. (Master's thesis). Retrieved from https://scholarcommons.sc.edu/etd/6943