An improved attribute importance degree algorithm based on rough set

In order to improve the efficiency of attribute importance degree algorithm, the thesis brings forward an new and efficient algorithm. It sorts the decision table by the radix sorting based on distribution and statistics at first and then calculates the equivalence class, so it reduces the time complexity of equivalence class algorithm from O(|C∥U|²) to O(|C∥U|). It is on the basis of the improved equivalence class algorithm that the improved attribute importance degree algorithm is presented. The analysis of an example and the result of experiments prove that the result of calculating attribute importance degree by the traditional and the improved attribute importance degree algorithm are completely same, but the improved attribute importance degree algorithm brought forward by the thesis is more efficient.