Distributed Algorithm for Computing Formal Concepts Using Map-Reduce Framework

Searching for interesting patterns in binary matrices plays an important role in data mining and, in particular, in formal concept analysis and related disciplines. Several algorithms for computing particular patterns represented by maximal rectangles in binary matrices were proposed but their major drawback is their computational complexity limiting their application on relatively small datasets. In this paper we introduce a scalable distributed algorithm for computing maximal rectangles that uses the map-reduce approach to data processing.

• Type: Conference
• Year: 2009
• Where: IDA
• Authors: Petr Krajca, Vilém Vychodil
• Keywords: Binary Matrices, Formal Concept Analysis, IDA 2009, Information Management, Maximal Rectangles