Definition:

Group features based on their dependencies.

Variations:

Agglomerative Hierarchical Clustering (AHC):

A "bottom up" approach where each observation starts in its own cluster, and pairs of clusters are merged as one moves up the hierarchy.

Divisive Hierarchical Clustering (DHC):

A "top down" approach where all observations start in one cluster, and splits are performed recursively as one moves down the hierarchy.

Priority Order:

Group > Extraction > Categorize

Inputs:

• Source Code

Outputs:

• Feature tree
• Feature clusters
• Dendrogram tree.

Examples:

• (Weston et al. 2009)
• (Kelly et al. 2011)
• (Chen et al. 2005)
• (Bécan 2013)
• (Nöbauer et al. 2014b)
• (Rubin et al. 2012)
• (Damasevicius et al. 2012)
• (Eyal-Salman et al. 2014)
• (Alves et al. 2008)

Tools:

• Cluster 3.0;
• Pycluster;
• C Clustering Library; More information here

Related Techniques:

• Formal Concept Analysis

Recommended situations:

Clustering is highly recommended in products that possesses high level of dependencies between feature implementations. Besides that, a good documentation is not required when applying this technique.

Not Recommended situations:

As an static analysis technique, clustering may be unable to find all elements related to the same feature when applied in source code where the implementation of a feature is spread across several modules.