Изграждане на съставни индикатори с клъстерен анализ
Construction of composite indicators with cluster analysis

Author(s): Donka Keskinova
Subject(s): Social Sciences, Sociology, Applied Sociology, Evaluation research
Published by: Пловдивски университет »Паисий Хилендарски«
Keywords: cluster analysis; typology; composite Indicators; Ward’s method; binary variables
Summary/Abstract: The monograph is a result of the doctoral study on "Problems and Approaches in the Application of Cluster Analysis for the Construction of Composite Indicators in Sociology". For the first time in Bulgaria this book explores the possible solutions in the methodology of the classical methods of cluster analysis, their consequences for the results of the analysis; additionally, it presents and systematizes the reasons to prefer one solution to another. This provides an opportunity to construct an adequate strategy for the application of the analysis so as to solve a specific research task in the social sciences.Two types of classifications in sociology are justified - based on a typological composite indicator and empirical typology. Approaches have been developed to construct a composite indicator (variable) through cluster analysis that allow for a more effective reduction of information in empirical sociological research and its use to reveal relations and dependencies.The research in the monograph also contributes to the discussion of issues in cluster analysis methodology as they relate to the efficiency of clustering methods, Ward method's proximity measures, the proximity measures used for binary variables; the use of standardization of variables; how to sort the cases in the database, among others.

• Details
• Contents