A Bootstrap Test Using Maximum Likelihood Ratio Statistics to Check the Similarity of Two 3-Dimensionally Oriented Data Samples1

    loading  Checking for direct PDF access through Ovid

Abstract

Comparing three-dimensionally oriented datasets is a problem encountered in various branches of earth science. A simple statistical tool for the comparison of two 3-dimensionally oriented datasets using the bootstrap method in line with the usual nonparametric permutation test is described here. This bootstrap test involves the estimation of maximum likelihood ratio statistic for properly constructed joint frequency tables of the datasets to be compared. This test does not use asymptotic result and will work well even for small sample sizes. Also this test does not make any specific distributional assumptions.

Related Topics

    loading  Loading Related Articles