摘要
提出了一个伸缩性好的空间决策树分类算法,在分类时既考虑待分类对象的非空间属性,又考虑其空间邻接对象的属性对其分类的影响。该算法没有训练数据库需存储于内存的限制,对训练库也没有记录个数及属性个数的限制,能生成简练、精确的决策树。
A flexible spatial decision tree classification algorithm is proposed. This algorithm takes not only attributes of classifying object but also attributes of their neighbor objects into account while classifying spatial objects. Having no restriction of main memory and record number and attribute number of training database, the algorithm can output practice and accurate decision tree.
出处
《计算机工程》
EI
CAS
CSCD
北大核心
2005年第7期93-95,共3页
Computer Engineering
基金
云南省自然科学基金资助项目(2002F0013M)
关键词
空间分类挖掘
算法的可伸缩性
属性列表
决策树分类
Spatial classification data mining
Flexible problem of algorithm
Attribute list
Decision tree classification