KDD-SC: Subspace Clustering Extensions for Knowledge Discovery Frameworks

TL;DR

KDD-SC is a versatile extension for popular knowledge discovery frameworks that adds comprehensive subspace clustering capabilities, including algorithms, evaluation, and visualization tools, facilitating high-dimensional data analysis.

Contribution

It introduces a framework-agnostic, plugin-based subspace clustering extension compatible with KNIME, RapidMiner, and WEKA, enhancing their high-dimensional data analysis functionalities.

Findings

Supports multiple subspace clustering algorithms

Provides integrated evaluation and visualization tools

Easily extendable to various KDD frameworks

Abstract

Analyzing high dimensional data is a challenging task. For these data it is known that traditional clustering algorithms fail to detect meaningful patterns. As a solution, subspace clustering techniques have been introduced. They analyze arbitrary subspace projections of the data to detect clustering structures. In this paper, we present our subspace clustering extension for KDD frameworks, termed KDD-SC. In contrast to existing subspace clustering toolkits, our solution neither is a standalone product nor is it tightly coupled to a specific KDD framework. Our extension is realized by a common codebase and easy-to-use plugins for three of the most popular KDD frameworks, namely KNIME, RapidMiner, and WEKA. KDD-SC extends these frameworks such that they offer a wide range of different subspace clustering functionalities. It provides a multitude of algorithms, data generators, evaluation measures, and visualization techniques specifically designed for subspace clustering. These functionalities integrate seamlessly with the frameworks' existing features such that they can be flexibly combined. KDD-SC is publicly available on our website.