Model proporcjonalnego hazardu Coxa przy różnych sposobach kodowania zmiennych
The Cox Proportional Hazard Model for Different Methods of Encryption of Variables

Summary/Abstract: Methods of survival analysis are more and more often used in analysis of social and economic occurrences. Due to lack of distributional information regarding the random variable, much attention is put on non-parametric or semi-parametric models. They are more and more commonly used for analysis of occurrences different than life expectancy. The condition of use of models of survival analysis is appropriate database that makes possible estimation of duration time of defined state for particular elements of analysed population. They are usually retrospective analyses with use of records. The example of such database is unemployment records. The article presents results of analysis of influence of encryption of variables on estimation of parameters of the Cox proportional hazard model and their interpretation. The authors also presented correlation between parameters of the model estimated for the data encrypted in two ways. The cohort consisted of the unemployed persons unregistered in specific period. Sub-clusters were allocated with respect to age that is a determinant of period of waiting for a job.

• Details
• Contents