摘要
针对传统的决策树生成算法的不足，提出了两种改进算法．实例说明，改进算法具有更好的优化效果，且证明了传统算法是改进算法２ 的特例．把不确定信息以条件概率的形式引入决策表，提出了条件决策表的概念及条件决策树的构造算法，拓宽了决策表的应用范围，使用决策表作知识库、决策树生成算法作推理机。
Two decision tree generation algorithms with improved efficiency are proposed. By using conditional probability, the fuzzy information is added into the original decision table to form a conditional decision table extending its range of applicability. The frame of a test & diagnosis expert system can thus be constructed. With the knowledge of the decision table, the reasoning mechanism can be expressed through the decision tree.
出处
《西安交通大学学报》
EI
CAS
CSCD
北大核心
1999年第10期43-47,共5页
Journal of Xi'an Jiaotong University
基金
国家自然科学基金
关键词
决策表
决策树
信息熵
条件决策表
生成算法
decision table
decision tree
information entropy
conditional decision table
test optimization
