Abstract
A rule-based classifier learns rules from a set of training data instances with assigned class labels and then uses those rules to assign a class label for a new incoming data instance. To accommodate data imperfections, a probabilistic relational model would represent the attributes by probabilistic functions. One extension to this model uses belief functions instead. Such an approach can represent a wider range of data imperfections. However, the task of extracting frequent patterns and rules from such a "belief theoretic" relational database has to overcome a potentially enormous computational burden. In this work, we present a data structure that is an alternate representation of a belief theoretic relational database. We then develop efficient algorithms to query for belief of itemsets, extract frequent itemsets and generate corresponding association rules from this representation. This set of rules is then used as the basis on which an unknown data instance, whose attributes are represented via belief functions, is classified. These algorithms are tested on a data set collected from a testbed that mimics an airport threat detection and classification scenario where both data attributes and threat class labels may possess imperfections.
| Original language | English |
|---|---|
| Title of host publication | 2005 7th International Conference on Information Fusion, FUSION |
| Pages | 661-668 |
| Number of pages | 8 |
| Volume | 1 |
| DOIs | |
| State | Published - Dec 1 2005 |
| Event | 2005 7th International Conference on Information Fusion, FUSION - Philadelphia, PA, United States Duration: Jul 25 2005 → Jul 28 2005 |
Other
| Other | 2005 7th International Conference on Information Fusion, FUSION |
|---|---|
| Country | United States |
| City | Philadelphia, PA |
| Period | 7/25/05 → 7/28/05 |
Fingerprint
Keywords
- Association rules
- Classification
- Data ambiguities
- Data imperfections
- Data mining
- Dempster-Shafer belief theory
ASJC Scopus subject areas
- Engineering(all)
Cite this
Rule mining and classification in imperfect databases. / Hewawasam, K. K R G K; Premaratne, Kamal; Subasingha, S. P.; Shyu, Mei-Ling.
2005 7th International Conference on Information Fusion, FUSION. Vol. 1 2005. p. 661-668 1591917.Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Rule mining and classification in imperfect databases
AU - Hewawasam, K. K R G K
AU - Premaratne, Kamal
AU - Subasingha, S. P.
AU - Shyu, Mei-Ling
PY - 2005/12/1
Y1 - 2005/12/1
N2 - A rule-based classifier learns rules from a set of training data instances with assigned class labels and then uses those rules to assign a class label for a new incoming data instance. To accommodate data imperfections, a probabilistic relational model would represent the attributes by probabilistic functions. One extension to this model uses belief functions instead. Such an approach can represent a wider range of data imperfections. However, the task of extracting frequent patterns and rules from such a "belief theoretic" relational database has to overcome a potentially enormous computational burden. In this work, we present a data structure that is an alternate representation of a belief theoretic relational database. We then develop efficient algorithms to query for belief of itemsets, extract frequent itemsets and generate corresponding association rules from this representation. This set of rules is then used as the basis on which an unknown data instance, whose attributes are represented via belief functions, is classified. These algorithms are tested on a data set collected from a testbed that mimics an airport threat detection and classification scenario where both data attributes and threat class labels may possess imperfections.
AB - A rule-based classifier learns rules from a set of training data instances with assigned class labels and then uses those rules to assign a class label for a new incoming data instance. To accommodate data imperfections, a probabilistic relational model would represent the attributes by probabilistic functions. One extension to this model uses belief functions instead. Such an approach can represent a wider range of data imperfections. However, the task of extracting frequent patterns and rules from such a "belief theoretic" relational database has to overcome a potentially enormous computational burden. In this work, we present a data structure that is an alternate representation of a belief theoretic relational database. We then develop efficient algorithms to query for belief of itemsets, extract frequent itemsets and generate corresponding association rules from this representation. This set of rules is then used as the basis on which an unknown data instance, whose attributes are represented via belief functions, is classified. These algorithms are tested on a data set collected from a testbed that mimics an airport threat detection and classification scenario where both data attributes and threat class labels may possess imperfections.
KW - Association rules
KW - Classification
KW - Data ambiguities
KW - Data imperfections
KW - Data mining
KW - Dempster-Shafer belief theory
UR - http://www.scopus.com/inward/record.url?scp=33847138707&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=33847138707&partnerID=8YFLogxK
U2 - 10.1109/ICIF.2005.1591917
DO - 10.1109/ICIF.2005.1591917
M3 - Conference contribution
AN - SCOPUS:33847138707
SN - 0780392868
SN - 9780780392861
VL - 1
SP - 661
EP - 668
BT - 2005 7th International Conference on Information Fusion, FUSION
ER -