Research on Classification Algorithm of One Improved Decision Tree
WU Bixia · Journal of Chengdu University · 2011
The traditional decision tree ID3 algorithm has many disadvantages,such as low computational efficiency,variety bias and so on.A scheme was proposed to improve attribute value vacancy,multi-valued attribute selection and attribute selection criteria and weighted entropy and simplified entropy were introduced into decision tree algorithm to improve traditional ID3 algorithm.The experimental results show that the improved algorithm has more excellent performance than present extensive use ID3 algorithm in terms of overall performance.