When the empirical data is incomplete (truncated or censored), raw empirical estimators will not produce good results. In this scenario, there are two techniques available to determine the distribution function based on the data. The *Kaplan-Meier product limit estimator* can be used to generate a survival distribution function. The *Nelson-Aalen estimator* can be used to generate a cumulative hazard rate function. The Kaplan-Meier estimator is given by:

where is the risk set at time and is the number of observations from the random event whose distribution you are trying to estimate. For example, if the random event you are interested in is death, then could be the number of life insurance policy holders immediately prior to the first death, and would be the number of observed deaths at the time of the first death (you can have simultaneous deaths). The key to dealing with problems that use this estimator is to understand how changes with respect to censoring or truncation. If a person withdraws from the life insurance policy, this decreases but this is not a death, so it does not contribute to . If new members join at time , they are not part of the risk set until time .

If the data is censored past a certain point, you can assume an exponential distribution for the censored portion. Suppose observations past are censored. If you know the value of you can solve for using .

The Nelson-Aalen cumulative hazard rate estimator is given by:

You can use this to get a survival function: