Jeszcze raz o skutkach subiektywizmu w analizie wielowymiarowej
The effects of subjectivism in multivariate analysis revisited

Summary/Abstract: This article provides the completion of the in-press article The story about consequences of subjectivism in multivariate analysis, by delivering information on the effects of subjectivism in choosing the method of classification of complex objects. The obvious effect of subjectivism is the difference in results of the classification. Conducted research brings the following conclusions: 1) the difference in classification results depends on the number of clusters: the more the clusters the bigger the difference; 2) the effects of subjectivism are the most apparent in case of choosing the method of classification (Wroclaw numerical taxonomy tends to differentiate the results less than K-Means Method), less noticeable are the effects of subjectivism in choosing the type of distance measure, finally – occurring rarely than in case of choosing the distance measure – the effects of subjectivism when choosing the method of normalization of the characteristics of complex objects.

• Details
• Contents