Analysis and Improvement of Classification Based on Multiple Association Rules

DOI

10.2991/icaita-18.2018.40How to use a DOI?

Keywords

classification based on Multiple Association Rules; support threshold; FP tree; classification association rules

Abstract

An excellent algorithm is critical to classify data. The algorithm of Classification based on Multiple Association Rules combines association rules with classification. The limitations of traditional CMAR algorithm are analyzed and an improved algorithm is presented. In the new algorithm, each attribute of sample data has a weight according to its importance. The weight of every attribute revises the basic support threshold. So the unique support threshold of each attribute is calculated. The support of each attribute value is compared with the unique support threshold of the attribute. If the support of an attribute value is greater than the unique support threshold, a set NF-List will contain this attribute value. A new FP tree is built on the basis of the set NF-List. The new FP tree is traversed to create classification association rules by Apriori. Experiments show that the improved CMAR algorithm can generate more association rules about important attributes, but not the ordinary rules in the traditional CMAR algorithm.

Copyright

© 2018, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Jing Lin
PY  - 2018/03
DA  - 2018/03
TI  - Analysis and Improvement of Classification Based on Multiple Association Rules
BT  - Proceedings of the 2018 2nd International Conference on Artificial Intelligence: Technologies and Applications (ICAITA 2018)
PB  - Atlantis Press
SP  - 158
EP  - 161
SN  - 1951-6851
UR  - https://doi.org/10.2991/icaita-18.2018.40
DO  - 10.2991/icaita-18.2018.40
ID  - Lin2018/03
ER  -