A Two Step Parallel Discretization Algorithm Based on Dynamic Clustering

A two-step discretization algorithm by dynamic clustering based on Rough set is proposed. The algorithm first discretization for decision table using dynamic clustering algorithm, then discrete again using cut importance discretization algorithms, and obtain the final cut sets, because the dynamic clustering algorithm processing speed is quickly and it impel a lot of breakpoint be screened, thus the operation efficiency of algorithm is increase notable. Finally, simulation results show that the proposed algorithm is correct and high efficiency.