PEMODELAN COPULA CLAYTON UNTUK PREDIKSI KLAIM PADA DATA LONGITUDINAL DENGAN EXCESS ZEROS

Anaviroh Anaviroh (Universitas Islam Darul ‘Ulum Lamongan)
Adhitya Ronnie Effendie (Universitas Gadjah Mada)

Unisda Journal of Mathematics and Computer Science (UJMC)
Vol 1 No 01 (2015): Unisda Journal of Mathematics and Computer Science

This papper discuss about longitudinal data models of claim counts withexcess-zeros, in which time-dependence of the claim counts is modeled by using acopula function. The copula approach extensively to model the serial dependence ofthe claim counts in car insurance, to model this serial dependence of the claimcounts (between the history and future claims). The maximum likelihood is appliedto estimate the parameters of the discrete copula model. A two-step procedure isproposed to estimate the parameters and predict the claim counts of the next periodusing the estimated parameters.