Search

The steady growth in the size of data has encouraged the emergence of advanced main memory trie-based data structures. Concurrently, more acute knowledge extraction techniques are devised for the discovery of compact and lossless knowledge formally expressed by generic bases. In this paper, we present an approach for deriving generic bases of association rules. Using this approach, we construct small partially ordered sub-structures. Then, these ordered sub-structures are parsed to derive, in a straightforward manner, local generic association bases. Finally, local bases are merged to generate the global one. Extensive experiments carried out essentially showed that the proposed data structure allows to generate a more compact representation of an extraction context comparatively to existing approaches in literature. [ABSTRACT FROM AUTHOR]