Analysis of estimation methods for the extremal index

Many datasets present time-dependent variation and short-term clusteringwithin extreme values. The extremal index is a primary measure to evaluate clusteringof high values in a stationary sequence. Estimation procedures are based on the choiceof a threshold and/or a declustering parameter or a block size. Here we revise severaldierent methods and compare them through simulation. In particular, we will seethat a recent declustering methodology may be useful for the popular runs estimatorand for a new estimator that works under the validation of a local dependence condition. An application to real data is also presented