|
|||||
|
|
| 基于聚类的决策树连续属性离散化改进算法 | |
| 引用本文： | 周锐,胡学钢. 基于聚类的决策树连续属性离散化改进算法[J]. 微计算机信息, 2011, 0(6) |
| 作者姓名： | 周锐 胡学钢 |
| 作者单位： | 合肥工业大学计算机与信息学院;安徽广播电视大学; |
| 摘 要： | 基于信息熵的二元分割算法离散连续属性,在对连续属性较多,数据量较大的数据集进行分析预测中,存在不足。实验表明,在决策树算法中结合改进后的k-means算法作为连续属性离散化算法,在连续属性较多的数据实例中可以构造出更好的决策树。 |
| 关 键 词： | 连续属性 离散化 改进算法 聚类 K-means算法 |
A Improved Algorithm for Discretization of Continuously-valued Attribute in Decision Tree based on Clustering |
|
| ZHOU Rui HU Xue-gang. A Improved Algorithm for Discretization of Continuously-valued Attribute in Decision Tree based on Clustering[J]. Control & Automation, 2011, 0(6) | |
| Authors: | ZHOU Rui HU Xue-gang |
| Affiliation: | ZHOU Rui HU Xue-gang(chool of Computer and Information,Hefei University of Technology,Hefei 230009,China)(Anhui Radio & TV University,Hefei 230022,China) |
| Abstract: | When the large data-set had many continuously-valued attribute,it was insufficient that the bi-interval discretization method based on information entropy discreted continuously-valued attribute.Experimental results show that,using improved k-means algorithm for discretization of continuous attributes can be constructed in the many continuously-valued attribute instance focus on a better decision tree. |
| Keywords: | Continuously-valued Attribute Discretization Improved Algorithm clustering algorithm K-means algorithm |
| 本文献已被 CNKI 等数据库收录！ | |