Analysis of Multi-Choice Questionnaires
through Self-Organizing Maps

Anssi Lensu and Pasi Koikkalainen
Department of Mathematics, University of Jyväskylä
P.O.Box 35, FIN-40351 Jyväskylä, Finland

Abstract:

This paper describes how Self-Organizing Maps can be used to analyse multi-choice gallups. In this method, the use of a single SOM for all available data is replaced with the use of multiple SOMs trained with subsets of gallup questions. The subgroupings located from these maps are then used to train a new concluding SOM that is more readable than any single SOM analysis would be.

• Introduction
• Algorithm for the two-Phase analysis
• Preprocessing and data field categorization
  • Coding via fuzzy memberships
  • Motivation for categorization of answers
• Combining categorical analyses
  • Group membership calculation
• Making conclusions and fitness estimation
• Example data sets
  • Training and linquistic conclusions
• Conclusions
• References
• About this document ...

Back to SECI publications