Cluster Bombe

This program is designed to find morphologically similar names in a document, cluster them together, and then find social groups (families, farms, etc.) that have a higher than random chance of being composed of elements from that cluster. By taking a nameÆs construction as a proxy for familial relationships, we hope to find groups that are likely to be related.

• The original specification and a more technical overview.
• A translation of the first.
• The program itself.
• The included README