A rule-set based decision tree for classification

• 1、School of Information and Communication Engineering, Beijing University of Posts and Telecommunications

Abstract：Decision tree is one of the most useful tools for solving the classification problems. This method requires building a decision tree to model the process of classification. The tree can be applied to the data instance so as to produce the result of classification after it has been built. Most of the decision trees are the training based on the data set. The disadvantage of this method is that once the tree is built, it's difficult to modify it to suit different decision making situation because of the large scale of data set. This paper proposes a decision tree training method based on rule set. The method includes two steps: 1 - using ID3 algorithm to transform the data set into the rule set; 2- training the decision tree based on the rule set using three attribute selection criteria AO, AD and MVD. In the simulation results, it can be seen that the decision tree trained based on the rule set is less complex compared with one trained by ID3 algorithm.

Keywords： data mining decision tree rule set attribute selection criterion