Comparison of Selected Criteria for Determination of the Measure of Depth of an Observation in a Two-dimensional Sample

The measure of observation depth in multidimensional samples, introduced into statistical practice by Tukey, has become a new tool for data analysis. It is a proposed method for determining multi-dimensional positional statistics, particularly in the analysis of non-typical data with outstanding observations. Applying a rule of depth helps to overcome the difficulties associated with sorting multidimensional observations. The notion of data depth has been intensively developed by many scholars, and, consequently, various criteria of the measurement of observation depth in a multidimensional samples may be found in literature. This paper contains a comparison of selected criteria of the measurement of observation depth in a two-dimensional case.

W artykule dokonano porównania wybranych kryteriów wyznaczania miary zanurzania obserwacji w statystycznej próbie dwuwymiarowej. Wnioski dotyczące porównania tych kryteriów wyciągnięto na podstawie własnych badań empirycznych na próbach dwuwymiarowych.