[1]戴 南,吉根林.分布式决策树算法研究与实现[J].南京师范大学学报(工程技术版),2005,05(04):046-48.
　DA I Nan J IGenlin.Research and Implementation of ID3 Based on Distributed Database System[J].Journal of Nanjing Normal University(Engineering and Technology),2005,05(04):046-48.
点击复制

分布式决策树算法研究与实现
分享到：

南京师范大学学报（工程技术版）[ISSN:1006-6977/CN:61-1281/TN]

卷:
05卷
期数:
2005年04期
页码:
046-48
栏目:
出版日期:
2005-12-30

文章信息/Info

Title:
Research and Implementation of ID3 Based on Distributed Database System
作者:
戴  南1;  2;  吉根林1;  2
1. 南京师范大学数学与计算机科学学院, 江苏南京210097;
2. 苏州大学江苏省计算机系统信息处理重点实验室, 江苏苏州215006
Author(s):
DA I Nan 1;  2;  J IGenlin 1;  2
1. School of M ath em atics and Computer Science, Nan jing Norm alU niversity, Jiangsu Nan jing 210097, China;
2. Key Lab of Computer In formation Process ing of J iangsu Province, S oochow Unnvers ity, Jiangsu Suzhou 215006, China
关键词:
分类;  决策树;  分布式决策树
Keywords:
classify;  dec ision tree;  d istr ibu ted dec ision tree
分类号:
TP301.6
摘要:
提出了一种基于分布多库环境下的决策树生成算法DDTA(D istributed D ecision Tree A lgorithm).该算法使用基于信息熵增益的思想分割各个分布的、同构训练样本集,各分布站点利用服务器传来的分割属性分割自己的样本集,服务器则通过对所有分布站点传来的信息计算各个属性的信息熵增益得到分割属性.实验表明DDTA算法能对分布同构样本集进行有效决策树挖掘,分布多库环境下生成的决策树是正确的.与算法INDUS相比,该算法的通信代价小.
Abstract:
A new dec ision tree algor ithm DDTA( D istributed Dec is ion Tree A lgor ithm ) based on distributed data re positor ies is presented in th is paper. The algor ithm divides each distributed and isomo rph ic da ta set uses w ith the idea of in fo rm ational entropy increase. E ach distributional site d iv ides its own data reposito ry w ith the d iv id ing properties transm itted by the se rver, and the server obta ins the dividing properties by ca lcu la ting the in fo rm ational entropy in crease of var ious properties w ith inform ation transm itted from a ll the distr ibuted sites. The expe rim en t show s that DD TA a lgo rithm is e ffective in excavating d istributiona lly isom orphic data repository w ith a dec ision tree, and that the de cision tree genera ted in the env ironm ent of d istributional mu lti-data repositor ies is co rrect. Com pared w ith the a lgo rithm INDUS, the algorithm has less cost in communication.

相似文献/References:

[1]杨杨,刘会东.一种基于成对约束的特征选择改进算法[J].南京师范大学学报(工程技术版),2011,11(01):056.
　Yang Yang,Liu Huidong.An Improved Algorithm for Feature Selection Based on Pairwise Constraint[J].Journal of Nanjing Normal University(Engineering and Technology),2011,11(04):056.

备注/Memo

备注/Memo:
基金项目: 江苏省重点实验室开放基金资助项目( KJS03064) .
作者简介: 戴  南( 1979-) , 女, 助教, 主要从事数据挖掘方向的教学与研究. E-m ail:dainan@ njnu. edu. cn
更新日期/Last Update: 2013-04-29