Clustering Mixed-Type Data
Emmanouil D. Pratsinakis, Zacharenia Kyrana, Nikolaos Papafilippou, Lefkothea Karapetsi, Christos A. Dordas, Angelos Markos · 2025
A multidimensional and multivariate structure with mixed-type data allows researchers to use various statistical approaches and data clustering techniques. The choice of clustering method used can have an impact on the results obtained In this study, the Partitioning Clustering (k-means) and Hierarchical Cluster Analysis methods were compared. The main objective of this study was to apply and compare these methods in segmenting the data set into groups-clusters. The results of the study showed that there are several analysis strategies for clustering mixed-type data, based on data coding and the chosen measurement scale for input variables. Remarkably, Python delivered results over 100% faster than SPSS. In conclusion, it emerged that the results of the classification depend on the coding strategy and the selection of the measurement scale of the variables to be used in the analysis.