Kredibilitní modely pro škodní frekvenci

Abstract

The work deals with estimation of unknown risk parameters of a driver. Risk parameter indicates how many times more accidents we may expect from this driver compared with the average insurance group to which the driver is assigned according to the precarious classification. Risk parameter is a random variable, which is depending on the damage share of this driver. The second and third chapters describe the derivation of the estimates that minimize the quadratic and exponential loss function. It also compares the level of these estimates and the convergence rate. In chapter four there are several simulations performed and tables with estimates for any specific data created. 1