Ocena wpływu wartości stałej Minkowskiego na możliwość identyfikacji struktury grupowej danych o wysokim wy-miarze
The assessment of impact value of Minkowski’s constant for the possibility of group structure identification in high dimensional data

Summary/Abstract: An important decision in the analysis of the variability of units in themultidi-mensional space, is the choice of the measurement of distance which is accurate for a given problem. This choice is of particular importance, when we have data sets which are described by hundreds of features. In the empirical studies, the most used measure of distance is the exponential metric measure. When the units are described by a very large number of features, an appropriate choice of the Minkowski’s constant level is important because has to affect the properties of the exponential metrics. With the increase of dimensionality, the properties of the metrics may change. The aim of this paper is to estimate the influence of the Minkowski’s constant and high dimensional space on the group structure which may be obtained. Based on simulation studies the author of this paper shows that the high dimension of the space application of fractional exponential metrics affects the ability to identify the group structure.

• Details
• Contents