Overview
One of the primary challenge for analyzing survival data is the issue of censored data. Censored data is data that is collected that is incomplete in some manner.
As a example, an insurance company collects information on the death events of people that they insure. Some customers buy life insurance, but then exit the policy prior to death. The information collected on this individual is incomplete, it is censored. However, the information that the individual had not died prior to exiting the policy contains useful information, it gives information as the likelihood of dying prior to the age of the individual when she exited the policy.
Methods
- Mortality Tables - are a simple way to try to fit a survival function to data. It often requires more data to be accurate than other methods.
- Fitting a Constant Hazard
- Fitting to Age
- Fitting to Factors