Application of Cluster Analysis When Studying Relationships Between Social Agents (Typical Problems and Approaches) Cover Image
  • Price 3.50 €

Приложение на клъстерния анализ при из-следване на отношения между социални дейци (типични проблеми и подходи
Application of Cluster Analysis When Studying Relationships Between Social Agents (Typical Problems and Approaches)

Author(s): Donka Keskinova, Ivan Tchalakov
Subject(s): Social Sciences
Published by: Институтът по философия и социология при БАН
Keywords: contemporary methods of data analysis; statistical analysis of sociological data; cluster analysis of nominal sociological data; multiple correspondence analysis on nominal sociological data; socio-technical networks; social network approach

Summary/Abstract: The paper presents the authors’ experience in application of cluster analysis of nominal data about relationships between social agents. It describes some typical problems that occur when clustering different type of variables and the solutions found by the authors. We have in mind the well know in the literature joint use of cluster analysis and multiple correspondence analysis (Lebart in Greenacre & Blasius 1994) on nominal sociological data with large number of cases and variables describing them. In the paper cluster analysis is shown as preliminary step to multiple correspondence analysis. The introductory part briefly describes the cluster analysis, while second part provides the necessary information about the main methods of clustering, stressing on hierarchical agglomerative methods, used by the authors. The third part considers some specific approaches for clustering different types of variables as illustration of different type of research problems.

  • Issue Year: 33/2001
  • Issue No: 3+4
  • Page Range: 36-59
  • Page Count: 24
  • Language: Bulgarian