The analysis and optimization of decision tree based on ID3 algorithm

Decision tree is one of the complex algorithms in the fields of pattern recognition and data mining, which is not only related to database, artificial intelligence and other disciplines, but has great theoretical research value. This paper used standard data set as original discrete experimental data, and the entropy and information gain of each attribute of the data was calculated to implement the classification of data. After traversing the structure of tree, the attributes corresponding to the information gain that reduces the maximum entropy were selected as optimal classification attribute for the generation of the decision tree. The simulation results show that the classification accuracy of ID3 decision tree was 6–7 percentage points higher than the other classification algorithms in this paper, meanwhile, the generation of the tree structure was apparently complex. In addition, this paper achieved an optimized structure of the decision tree, which is simplified to 5 node structure from the initial 19 nodes for improving the efficiency of the algorithm on the premise of ensuring low error rate which was at the same level as other classification algorithms.