Conditional L1 estimation for random coefficient integer-valued autoregressive processes

In this paper we study the integer-valued autoregressive model, which belongs to the class of thinning models with count data.We mainly focus on the random coefficient integer-valued autoregressive (RCINAR) model and propose a conditional least absolute deviation (CL1) method to estimate the parameters of the model. The asymptotic distribution of the CL1 estimator is investigated. The finite sample performance of the proposed estimator is evaluated through simulation, and is compared with that of conditional least squares (CL2) estimation method. Simulation results show that the proposed method is effective and robust against outliers