Blog
14/11/2023
In a previous article (read article: An alternative approach to the concept mapping method: clustering analysis on original distances), we discussed an alternative approach for concept mapping analysis. This consists in applying the cluster analysis directly to the raw distance matrix obtained from participants’ sorting of ideas rather than following the non-metric multidimensional scaling (nMDS). Projecting the data onto a two-dimensional space for cluster visualization is consequently done after the clusters are defined.
Based on a comparative analysis of different dimensionality reduction algorithms, we found that non-metric multidimensional scaling (nMDS) did not perform as well as other recent approaches for analysing non-metric distances. This is why, at Polygon, we are applying other non-linear dimensionality reduction methods to produce more valid concept maps.
We intend to publish the results of this comparative analysis in a scientific journal in order to share Polygon’s developments with the scientific community and to contribute to the advancement of concept mapping methods. These methodological developments are integrated into our CM* tool and will also be made available shortly in an open source Python package (read article: An open source package for concept mapping analysis).
Concept Mapping Tool
Participatory research
Research methods
Dimensionality reduction
Empirical data
Algorithm comparison
