iii
under a dependent censoring setting. Simulations are conducted to demonstrate the
method’s performance, and sensitivity analyses are performed to assess the impact of
the dependent censoring on the regression parameter estimates. In the last part of
methodology, we propose a method to test the existence and to identify the location
of a change-point in a hazard function. The application of our methodology to real
insurance companies’ survival data discloses important influence of the GLB Act on
insurance companies’ survival.