摘要
ID3算法是一种信息熵的决策树学习算法,把信息熵作为选择测试属性的标准,对训练实例集进行分类并构造决策树来预测如何由属性对整个实例空间进行划分。ID3算法对于相对小的数据集是很有效的,但对大型数据库而言,ID3算法无法处理。SLIQ分类算法使用了一些独特的技术,改进了学习的时间,同时在没有降低精确度的情况下,解决了对磁盘驻留大数据集的分类,具有更快的速度而且生成较小的树。
ID3 arithmetic is a kind of decision tree learned arithmetic based on informational entropy. Informational entropy is selected as standard of testing attribute. It can be used to classify and construct decision tree,and forecast how to compartmentalize the whole example space of disciplinary example concourse. ID3 is an effective arithmetic among the relative small data sets. On the contrary,ID3 is hardly to handle. SLIQ classified arithmetic makes use of some unique techniques,improves learning time,doesn't reduce precision,as well as solves the sort of large-scale data set of disk memory, and builds smaller tree as fast as possible .
出处
《计算机与现代化》
2005年第3期19-21,共3页
Computer and Modernization
关键词
ID3
SLIQ
分类器
MDL
决策树
ID3
supervised learning in quest
classified apparatus
minimum description length
decision tree
