Analysis of Association Rule Mining on Quantitative Concept Lattice

摘要

In the process of association rule mining on rough set, it is always needed to deleting the reduplicative rows or columns, so supports and confidences of association rules cannot be obtained accurately. While the Hasse diagram of quantitative concept lattice contains all the objects and attributes information, supports of nodes can be obtained visually from the lattice, and the vivid association rule mining can be realized. Association rule mining algorithm on quantitative concept lattice effectively avoids the combinatorial explosion problem existing in rough set. Confidences of rules can be obtained accurately via the supports of relative concept nodes, and it can also effectively avoid the problem of information loss existing in rough set reduction, thus the efficiency of association rule mining can be improved.