clusEvol: An R package for Cluster Evolution Analytics

Abstract

The paper proposes a new R package, named clusEvol, that introduces Cluster Evolution Analytics (CEA), a framework for advanced Exploratory Data Analysis and Unsupervised Learning. CEA studies the evolution of an object and its neighbors, identified via clustering algorithms, over time. It combines leave-one-out and plug-in principles, enabling “what if” scenarios by integrating current data into past datasets to explore temporal changes. The framework is demonstrated with a real dataset employing various clustering algorithms.

Keywords

• Exploratory data analysis
• Unsupervised learning
• Clustering
• Statistics