Comparison of the Use of Linkage in Cluster Integration With Path Analysis Approach

Abstract

This study aims to compare integrated clusters on various linkages (Ward, Complete, and Average linkage) with path analysis to classify the behavior of subservient to pay for mortgages. The data used are primary data from questionnaires filled out by Bank X customers, using purposive sampling. The measurement method used is the average score method. Liu et al. with the title “Enhanced path-based Clustering Algorithm” shows the results that path analysis with cluster integration gives results that outperform other cluster algorithms. So, this study uses the cluster integration with path analysis method by comparing the linkage. The results showed that cluster integration in the Ward linkage method with path analysis was the best method, by comparing the coefficient of determination. The variables of attitude, subjective norms, behavioral control, and intention to comply with paying explained the behavior of Bank X customers to comply with paying mortgages by 88.32%, while the remaining 11.68% was explained by variables outside the model.