摘要
决策树方法因结构简单、便于理解和具有较高的分类精度而在数据挖掘中被广泛采用,其规则生成算法实现对决策树规则的提取和化简。属性相关性分析的基本思想是计算某种度量,用于量化属性与给定概念的相关性。提出了一种基于属性相关性的c4.5决策树规则生成算法c-c4.5 ru les,可替代c4.5原有的规则生成算法。c-c4.5 ru les在对规则进行化简时充分考虑了属性之间的关联性,实验表明该算法在保持原有分类精度的前提下,能有效提高规则生成时的计算速度和效率。
Decision tree is used extensively as a classifier in data mining for its simple structure, wide comprehension and high classification precision. The rule generation algorithm is used to extract and simplify production rules from decision tree. The main idea of the attribute correlation is to quantify the correlation between attribute and concept. This paper proposes a new rule generation algorithm based on attribute correlation （c- c4. 5rules）. It takes full advantage of the correlation between attributes when simplifying the rules. Experiments show that the new algorithm speeds up the computation efficiently while keeping the original classification precision.
出处
《计算机仿真》
CSCD
2006年第12期90-92,103,共4页
Computer Simulation
关键词
决策树
规则生成
属性
相关性
Decision tree
Rule generation
Attributes
Correlation
