Title :
Algorithm with Weighted Attributes for Unresolved Exception in Decision Tree Induction Algorithm
Author :
Cao, Ying ; Zhang, Chun-hai
Author_Institution :
Dept. of Comput. Sci., Ocean Univ. of China, Qingdao, China
Abstract :
Decision tree classification algorithm provides a fast and effective classification method for datasets, and it calculates information gain of each attribute, and selects the attribute with the greatest information gain as the split. However, to the best of our knowledge, when we use the traditional decision tree algorithm to analyze data in real life, we will encounter some unusual situations. This paper presents a new algorithm that can solve problems which majority voting algorithm can not be overcome, and verification of the accuracy of decision tree induction algorithm is improved.
Keywords :
data analysis; data mining; decision trees; pattern classification; decision tree classification algorithm; decision tree induction algorithm; information gain; majority voting algorithm; real life data analysis; weighted attributes algorithm; Accuracy; Classification algorithms; Decision trees; Entropy; Gain measurement; Prediction algorithms; Training data; accuracy; dataset; decision tree; exception;
Conference_Titel :
Business Computing and Global Informatization (BCGIN), 2012 Second International Conference on
Conference_Location :
Shanghai
Print_ISBN :
978-1-4673-4469-2
DOI :
10.1109/BCGIN.2012.140