Agglomerating local patterns hierarchically with ALPHA

To increase the relevancy of local patterns discovered from noisy relations, it makes sense to formalize error-tolerance. Our starting point is to address the limitations of state-ofthe-art methods for this purpose. Some extractors perform an exhaustive search w.r.t. a declarative specification of error-tolerance. Nevertheless, their computational complexity prevents the discovery of large relevant patterns. Alpha is a 3-step method that (1) computes complete collections of closed patterns, possibly error-tolerant ones, from arbitrary n-ary relations, (2) enlarges them by hierarchical agglomeration, and (3) selects the relevant agglomerated patterns. Categories and Subject Descriptors: H.2.8 [Database Management]: Database Applications—Data mining General Terms: Algorithms

• Type: Conference
• Year: 2009
• Where: CIKM
• Authors: Loïc Cerf, Pierre-Nicolas Mougel, Jean-François Boulicaut
• Keywords: Arbitrary N-ary Relations, CIKM 2009, Database, Large Relevant Patterns, Relevant Agglomerated Patterns