|
Abstract:
|
This work proposes a statistical methodology to handle aggregate data. We consider individuals moving among a finite set of states and observations are only made at different discrete time points. Aggregate data arise when the only information available at these time points is the total number of individuals in every state. In such a case, the history of each individual is lost and the time spent in a given state is unknown. Aggregate data arises in many fields such as medical science, ecology, social science, reliability, etc. Using a simple data completion technique, we will show how the mean sojourn time can be approximated and we will study under which conditions such an approximation is adequate. Applications will be displayed using data from the literature and results will be compared with other approaches.
|
