Abstract:
Clustering techniques play a pivotal role in data analysis, facilitating the exploration and
organization of complex datasets into meaningful groups. This study proposes a novel
approach to the predefined clusters and hierarchical clustering (PECHC) approach, which
is intended for efficient clustering using predefined clusters. The primary objectives include
enhancing cluster quality balanced with efficiency. The study includes significant
parameters that involve the number of iterations, Silhouette Score, and Davies-Bouldin
Index to assess PECHC's effectiveness across several datasets. The results of the analysis
indicate that PECHC consistently achieved superior clustering performance and efficiency
relative to other methods. In terms of methodology, PECHC uses recursive plotting for
initial cluster estimation and principal component analysis (PCA) for dimensionality
reduction. PECHC shows high cluster separation, competitive clustering accuracy, and
efficient convergence on a variety of datasets.
