The inducing of classification rules based on parallel ECL

Inducing classification rules is a research area of machine learning and data mining that has received a lot of attention in recent years. ID3 and C4.5 are well-known algorithms for mining classification rules. However, the rules induced by ID3 are not optimal. Thus, a novel approach inducing classification rules in the extended concept lattice (ECL) improves the concept lattice by introducing the equivalent intent. The approach is superior to ID3 and C4.5. The new features of databases such as high-dimensionality and heterogeneity make parallel/distributed data mining a hot research domain. Under this circumstance, we propose parallel ECL for data mining from a large scale database. In this paper, inducing classification rules in parallel ECL is investigated both theoretically and experimentally.