Comparison of the Methods of Self-Organizing Maps and Multidimensional Scaling in Analysis of Estonian Emotion Concepts
Toomas Kirt, Ene Vainik · DSpace repository (University of Tartu) · 2007
Self-organizing map (SOM) and multidimensional scaling (MDS) are the methods of data analysis that reduce dimensionality of the input data and visualize the structure of multidimensional data by means of projection.Both methods are widely used in different research areas.In the studies of emotion vocabulary and other psycho-lexical surveys the MDS has been prevalent.In this paper both of the methods are introduced and as an illustration they are applied to a case study of Estonian emotion concepts.There is a need to introduce some new methods to the field because exploiting only one analytical tool may tend to reveal only specific properties of data and thus have an unwanted impact on the results.