摘要
针对传统ID3算法存在多值属性偏向及运算量大的问题,引入粗糙集思想,定义了条件属性的相容度。利用属性的相容度作为分裂数据集的标准,构造决策树,避免传统ID3算法中对数的计算及多值属性的偏向。在3个UCI公共数据集上进行仿真实验,结果表明提出的新属性选择方法具有更高的预测准确率。
Since the traditional ID3 algorithm has the problems of multi-value attribute deviation and heavy computational amount,the rough set thought is introduced,and the consistency degree of condition attributes is defined.The consistency degree of the attributes is taken as the standard splitting the dataset to construct the decision tree,which avoids the multi-value attribute deviation and logarithmic computation.The simulation experiment for the algorithm is carried out with the three UCI common datasets.The experimental results show that the new attribute selection algorithm has higher prediction accuracy than the traditional ID3 algorithm.
作者
王子京
刘毓
WANG Zijing;LIU Yu(School of Communications and Information Engineering,Xi’an University of Posts&Telecommunications,Xi’an 710121,China)
出处
《现代电子技术》
北大核心
2018年第23期9-12,共4页
Modern Electronics Technique
基金
陕西省工业攻关(2016GY-113)~~
关键词
数据挖掘
决策树
粗糙集
ID3算法
大数据
算法改进
data mining
decision tree
rough set
ID3 algorithm
big data
algorithm improvement
