This thesis provides a comprehensive introduction to the fundamental concepts of survival analysis and investigates the use of parametric regression models in analyzing time-to event data. Key quantities as well as non-parametric methods are discussed. Likelihood-based techniques for censored and truncated data are also developed. Parametric regression models assume an underlying failure distribution, of which the exponential, Weibull, log-normal and log-logistic distribution are the most common choices. Our focus lies on the accelerated failure time model (AFT), its model estimation, model selection and diagnostics. Finally, we apply the AFT models to evaluate real-world data and draw conclusions on the advantages and limitations of the parametric models.     «

This thesis provides a comprehensive introduction to the fundamental concepts of survival analysis and investigates the use of parametric regression models in analyzing time-to event data. Key quantities as well as non-parametric methods are discussed. Likelihood-based techniques for censored and truncated data are also developed. Parametric regression models assume an underlying failure distribution, of which the exponential, Weibull, log-normal and log-logistic distribution are the most common...     »