Survival analysis deals with the statistical analysis of the time point when a specific
event occurs. This event can be a person’s death, an engine failure, or the moment
a person leaves a shopping mall after entering. In many applications, the probability
of survival is influenced by one or more confounding factors, also called covariates. In
medical applications, this could be the age, gender, blood levels, or medical record of a
person. The proportional hazards model is one of the most famous models for data with
time-independent covariates. The subject of this thesis is to give a concise introduction
to the proportional hazards model. Maximum likelihood estimation can be tough with
the proportional hazards model because it is semi-parametric. This thesis presents a
solution using the partial likelihood methods developed by Cox for no ties and Breslow
for the case of few ties between the event times. Alongside that, testing and modeling
techniques commonly used in the field are introduced.
A data analysis is conducted to estimate the survival of breast cancer without recurrence,
employing the methods described in this thesis.
«
Survival analysis deals with the statistical analysis of the time point when a specific
event occurs. This event can be a person’s death, an engine failure, or the moment
a person leaves a shopping mall after entering. In many applications, the probability
of survival is influenced by one or more confounding factors, also called covariates. In
medical applications, this could be the age, gender, blood levels, or medical record of a
person. The proportional hazards model is one of the most f...
»