Wpływ redukcji liczby zmiennych na stabilność grupowania
The effect of reduction of variables to groups stability

Summary/Abstract: The paper examines the effect of formal methods of variables selection on groups stability. The stability criterion examines whether the groups that were formed as a result of using taxonomy methods actually exist in the data (the structure is stable) or they have come up by chance. As a formal method of selecting variables, factor analysis was used, while groups stability testing was performed using a method of approximately unbiased bootstrap probability on socio-economic data sets generated on the basis of data taken from Central Statistical Office. The results show that using factor analysis for reducing the number of variables can affect both the stability of grouping and the obtained groups structure.

• Details
• Contents