摘要
应用变精度粗糙集理论,提出了一种利用新的启发式函数构造决策树的方法。该方法以变精度粗糙集的分类质量的量度作为信息函数,对条件属性进行选择。和ID3算法比较,本方法充分考虑了属性间的依赖性和冗余性,尤其考虑了训练数据中的噪声数据,允许在构造决策树的过程中划入正域的实例类别存在一定的不一致性,可简化生成的决策树,提高决策树的泛化能力。
A new heuristic function to build decision trees based on variable precision rough set is proposed. The measure ofquahty ot classification acts as information function to select the condition attribute in this method. Compared with ID3 algorithm, dependency and redundancy between attributes are considered, especially noisy data of training sets. A certain inconsistency is allowed to exist in examples of the positive regions, so the decision trees is simplified and its extensive ability is improved.
出处
《计算机工程与设计》
CSCD
北大核心
2006年第17期3175-3177,共3页
Computer Engineering and Design
基金
江苏省高校自然科学基金项目(05KJB520048)。
