TY - GEN
T1 - Decision tree construction for genetic applications based on association rules
AU - Sami, Ashkan
AU - Takahashi, Makoto
PY - 2005/1/1
Y1 - 2005/1/1
N2 - A decision tree is an effective means of data classification from which rules can be both expressive and precise. However decision tree is only applicable in the applications that the data is expressed with attribute-value pairs. Since genetic data are not attribute-pairs, the only method that we know of to make decision-tree for them is based on a greedy graph-based data mining algorithm known as DT-GBI. Due to its greedy nature, some of the important rules may be missed. Even though recently some attempts to make the algorithm complete has been presented, the computational complexity of algorithm increased so much that is not appropriate for practical purposes. In this paper we present an approach to make decision tree for DNA based data, which basically uses regular association-rule algorithms. Thus it has computational complexity which is much more tractable. Our contribution in this paper is to convert DNA based problems to regular data mining methods and by this conversion we present a method that can be applied to all classes of classification based on association rules.
AB - A decision tree is an effective means of data classification from which rules can be both expressive and precise. However decision tree is only applicable in the applications that the data is expressed with attribute-value pairs. Since genetic data are not attribute-pairs, the only method that we know of to make decision-tree for them is based on a greedy graph-based data mining algorithm known as DT-GBI. Due to its greedy nature, some of the important rules may be missed. Even though recently some attempts to make the algorithm complete has been presented, the computational complexity of algorithm increased so much that is not appropriate for practical purposes. In this paper we present an approach to make decision tree for DNA based data, which basically uses regular association-rule algorithms. Thus it has computational complexity which is much more tractable. Our contribution in this paper is to convert DNA based problems to regular data mining methods and by this conversion we present a method that can be applied to all classes of classification based on association rules.
KW - DNA
KW - Decision support system
KW - Sequence estimation
UR - http://www.scopus.com/inward/record.url?scp=34249304483&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=34249304483&partnerID=8YFLogxK
U2 - 10.1109/TENCON.2005.301248
DO - 10.1109/TENCON.2005.301248
M3 - Conference contribution
AN - SCOPUS:34249304483
SN - 0780393112
SN - 9780780393110
T3 - IEEE Region 10 Annual International Conference, Proceedings/TENCON
BT - TENCON 2005 - 2005 IEEE Region 10 Conference
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - TENCON 2005 - 2005 IEEE Region 10 Conference
Y2 - 21 November 2005 through 24 November 2005
ER -