Abstract

During the last few years there have been radical changes in the computing architectures available for solving scientific problems. The introduction of parallel systems has opened new software avenues well-adapted to the solution of many of these. The present paper investigates the use of a parallel computer for the implementation of statistical algorithms that are based on processing a large number of samples taken from the data. Typical applications are the clustering of large data sets and the computation of robust regression coefficients. Some other possiblities are also mentioned.

Keywords

Resampling plans, Parallelization, Cluster analysis, Robust regression, Bootstrap.