Title :
Exploring an Improved Decision Tree Based Weights
Author :
Guo, Weizhao ; Yin, Jian ; Yang, Zhimin ; Yang, Xiaobo ; Huang, Li
Author_Institution :
Sch. of Inf. Sci. & Technol., Sun Yat-sen Univ., Guangzhou, China
Abstract :
Although decision tree learning has achieved great success in building classifier, most existing methods don´t pay attention to unequal weights between different instances from training and testing data sets. However, many real world data sets are imbalanced in nature. In this paper, we introduce a new improved decision tree based weights, which considers imbalanced weights between different instances, to address the class imbalanced problems. The proposed decision tree algorithm is simple and more effective in implementation than previous decision trees. Also, the new proposed algorithm will be compared with C4.5 (a novel decision tree algorithm) experimentally and the experiment results testify that our proposed algorithm outperforms C4.5 significantly, in terms of the improvement of the classification accuracy in UCI data sets.
Keywords :
decision making; learning (artificial intelligence); C4.5 algorithm; decision tree learning method; testing data set; training data set; weights decision tree; Classification algorithms; Classification tree analysis; Decision trees; Hospitals; Information science; Machine learning; Machine learning algorithms; Medical tests; Sun; Testing; cost-sensitive learning; imbalanced data; weight gain ratio; weight information entropy;
Conference_Titel :
Natural Computation, 2009. ICNC '09. Fifth International Conference on
Conference_Location :
Tianjin
Print_ISBN :
978-0-7695-3736-8
DOI :
10.1109/ICNC.2009.457