Statistická analýza cenzorovaných dat

Abstract

This thesis deal with the statistical analysis of censored data by means of simulation studies on the weibull, log-Normal, and gamma distributions. The simulations are carry out by 10,000 repetitions at different parameter settings, sample sizes (n) and censoring schemes. For to all distributions, the obtained results exhibit the accuracy of parameters for larger sample sizes (n), as well as the reliability of parameters to different combinations of the parameters values. The results support the idea that sample size is a key factor in the enhancement of the accuracy of the parameter estimates in the cases of censored data, detailed simulation results are provided including tables and figures illustrating the estimator’s behaviour under different conditions. These findings offer valuable insights for statistical inference involving censored data in various practical applications.This thesis deal with the statistical analysis of censored data by means of simulation studies on the weibull, log-Normal, and gamma distributions. The simulations are carry out by 10,000 repetitions at different parameter settings, sample sizes (n) and censoring schemes. For to all distributions, the obtained results exhibit the accuracy of parameters for larger sample sizes (n), as well as the reliability of parameters to different combinations of the parameters values. The results support the idea that sample size is a key factor in the enhancement of the accuracy of the parameter estimates in the cases of censored data, detailed simulation results are provided including tables and figures illustrating the estimator’s behaviour under different conditions. These findings offer valuable insights for statistical inference involving censored data in various practical applications.