Learning Decision Trees from Distributed Datasets

摘要

Decision trees are an important data mining tool with many applications.Like many classification techniques,decision trees process the entire database in order to produce a generalization of the data that can be used subsequently for classification.Distributed databases are not amenable to such a global approach to generalization.This paper describes architecture of decision trees induction from distributed datasets which includes configuration manager retrieval data from distributed data,pruning data,and partial decision trees and data integration.In retrieval data,we explore a general strategy for explores a general strategy transforming traditional machine learning algorithms into algorithms for learning from distributed data;then we devise a pruning algorithms to optimal the data retrieval;finally we integrate the distributed sub-result data into final decision trees.