Analysis of dependent data aggregated into intervals (Q2237827)

In some practical applications, data are reported as intervals. This paper, investigated the autocovariance-autocorrelation functions for interval-valued autoregressive series models. They derived the maximum likelihood estimators by exploiting the ideas of composite likelihood. The authors also studied the different internal distributions of the intervals themselves and, asymptotic properties of these estimators. Based on a simulation study, they showed that the new estimators perform considerably better than those obtained from the set-valued estimators.