Multi-relational decision tree algorithm: implementation and experiments

Date
2003-01-01
Authors
Atramentov, Anna
Journal Title
Journal ISSN
Volume Title
Publisher
Altmetrics
Authors
Research Projects
Organizational Units
Computer Science
Organizational Unit
Journal Issue
Series
Abstract

This work describes an efficient implementation (MRDTL-2) of the multi-relational decision tree learning algorithm based on the proposal by Knobbe et al. of the multi-relational data mining framework and implementation (MRDTL) by Leiva. Some simple techniques for speeding up the calculation of sufficient statistics for decision trees and related hypothesis classes from multi-relational data are shown in this work. Because missing values are fairly common in many real-world applications of data mining, this work also describes some simple techniques for dealing with missing values that were implemented in MRDTL-2. The results of the experiments with several real-world data sets from the KDD Cup 2001 data mining competition and PKDD 2001 discovery challenge are presented. The results of the experiments indicate that MRDTL-2 is competitive with the state-of-the-art algorithms for learning classifiers from relational databases.

Description
Keywords
Computer science
Citation
Source