Title :
A Method of Dealing with Numeric Attribute Based on Sample Distribution
Author :
Zhu, Weidong ; Lin, Yongmin
Author_Institution :
Sch. of Comput. & Inf. Technol., Beijing Jiaotong Univ., Beijing, China
Abstract :
In order to improve the predictive accuracy of inductive learning, a heavy analysis about the demerit of C4.5 in dealing with numeric attribute is given. By the method of estimating the probability distribution of the training samples, a new and simple method of dealing with numeric attribute is proposed in this paper. Experimental results of UCI data sets show that the proposed method has an excellent performance on accuracy issue and faster computing speed than C4.5 algorithm.
Keywords :
decision trees; learning by example; probability; C4.5 algorithm; inductive learning; numeric attribute; probability distribution estimation; sample distribution; Accuracy; Algorithm design and analysis; Classification algorithms; Clustering algorithms; Data mining; Entropy; Error analysis; distribution of training samples; entropy; numeric attribute;
Conference_Titel :
Information Technology and Computer Science (ITCS), 2010 Second International Conference on
Conference_Location :
Kiev
Print_ISBN :
978-1-4244-7293-2
Electronic_ISBN :
978-1-4244-7294-9
DOI :
10.1109/ITCS.2010.65
