Predictive compound risk models with dependence (Q2212152)

In the collective risk model for a given period of time, the sequence \(\{C_{j}\}_{j\in{\mathbb N}}\) of individual losses is i.i.d.\ and independent of the number of claims \(N\) and the total loss is defined as \(S=\sum_{j=1}^{N}C_{j}\). In the present paper, the authors consider a related and rather particular multiperiod model in which independence between individual losses and the number of claims is replaced by an assumption on the structure of \(E(S|N)\). For the total loss in a future period they propose a predictor which depends on claims experience in the past.