Red Trees: A relational decision tree algorithm in streams

Wei Hou

doi:10.1016/j.eswa.2010.02.096

Red Trees: A relational decision tree algorithm in streams

2010-09-01

7

PRE

AI

DOI：10.1016/j.eswa.2010.02.096

Original

Original request for help

Save

Abstract

En 中文

Classification of streaming data is one of the hottest research topics in data mining nowadays, many efforts had been dedicated to relative researches for the single stream. However, to the best of our knowledge, there is no counterpart algorithm for the multi-relational data streams up to now. In this paper, one data synopsis method, which is compatible with the scenario of multi-relational data streams, is introduced. Based on period sampling, this method could avoid multiple join operations at some extent. Pursuantly, an algorithm for constructing decision tree from multi-relational data streams, RedTrees, is proposed. Then, the declarative bias in RedTrees, JoinTree, which makes the pattern refinement more efficient, is discussed. The theoretical analysis and experiments prove its effectiveness and good efficiency. Crown Copyright (C) 2010 Published by Elsevier Ltd. All rights reserved.

Keywords:

Data ming

Multi-relational data streams

Decision tree

Period sampling

Red Trees: A relational decision tree algorithm in streams

Abstract

AI Summary

Journal

Organization

Cited Papers