A Novel Five-Step Data Mining Algorithm

Abstract

Based on the traditional data mining algorithm, a novel data mining algorithm is proposed. This algorithm consists of 5 steps: the first step, set the tree set; the second step, set the window third, subtree contribution; decision tree construction; the fourth step test, positive and negative examples set; the fifth step, expand the achievements window. The experimental study on open source data sets. The results showed that the five step proposed data mining method, not only can build a more concise decision tree, data mining and the accuracy is also higher than the traditional decision tree method.