用探测性的归纳学习方法从空间数据库发现知识

邸凯昌; 李德仁; 李德毅

doi:10.11834/jig.1999011222

数据挖掘与知识发现研究新进展 | 浏览量 : 0 下载量: 157 CSCD: 0

PDF
导出
分享
收藏
专辑

用探测性的归纳学习方法从空间数据库发现知识
Knowledge Discovery from Spatial Databases With Exploratory Inductive Learning
1999年4卷第11期页码：924
纸质出版：1999
DOI： 10.11834/jig.1999011222
稿件说明：

移动端阅览

邸凯昌, 李德仁, 李德毅. 用探测性的归纳学习方法从空间数据库发现知识[J]. 中国图象图形学报, 1999,4(11):924. DOI： 10.11834/jig.1999011222.

Knowledge Discovery from Spatial Databases With Exploratory Inductive Learning[J]. Journal of Image and Graphics, 1999, 4(11): 924. DOI： 10.11834/jig.1999011222.

摘要

将探测性数据分析，面向属性的归纳和rough集方法结合起来，形成了一种灵活通用的探测性归纳米学习方法EIL，可以从空间数据库中发现普遍知识，属性依赖，分类知识等多种知识，同时提出了和总结了多种生成空间数据库概念层次结构的方法用于归纳学习，用中国分省农业统计数据的发掘试验说明了EIL的可行性和有效性。

Abstract

An exploratory inductive learning approach is proposed to discover knowledge from spatial database. This approach integrates attribute oriented induction method with exploratory data analysis and Rough set methods. It is flexible and general purpose to discover general knowledge

attribute dependencies and classification knowledge and so on. In order to satisfy the requirements of background knowledge

several methods are proposed for generating concept hierarchies from data. An experiment on agricultural statistical data of China mainland shows that exploratory inductive learning approach is feasible and effective for spatial data mining.