Experimental Improvements to Regularity Clustering

Peters, Stephen Louis

Student Work

Experimental Improvements to Regularity Clustering

Public

Data clustering is an immensely powerful tool. The analysis of big data has led to many clustering techniques. Among these techniques is Regularity Clustering, a new technique based on Abel Prize winner Endre Szemerédi's Regularity Lemma. Regularity Clustering has been shown to outperform industry standard clustering techniques in many circumstances. In this report we present new methods of executing Regularity Clustering. Among these methods one, which we call the most recurring construction method, outperforms the standard Regularity Clustering method by a significant margin. We also present empirical evidence indicating when Regularity Clustering performs well.

This report represents the work of one or more WPI undergraduate students submitted to the faculty as evidence of completion of a degree requirement. WPI routinely publishes these reports on its website without editorial or peer review.

Creator