A comparative analysis of discretization algorithms for data mining

In this paper, four kinds of typical discretization algorithms were comparatively analyzed from two aspects using examples: one referred to the variable quality of classification and accuracy of approximation under different parameter, the other was the similarity degrees between reducted variable sets and the original variable set. On determination of reducted variable sets, the reduction was regarded as multi-objective optimization problem, which was solved by the genetic algorithm, and the optimal reducted variable sets were found through including degrees. Finally, the consistent conclusion on preference of discretization algorithms was gained.