TY - GEN
T1 - An ensemble approach to multi-label classification of textual data
AU - Kurach, Karol
AU - Pawłowski, Krzysztof
AU - Romaszko, Łukasz
AU - Tatjewski, Marcin
AU - Janusz, Andrzej
AU - Nguyen, Hung Son
N1 - Copyright:
Copyright 2013 Elsevier B.V., All rights reserved.
PY - 2012
Y1 - 2012
N2 - In this paper, we investigate different approaches to multi-label classification of textual data, with a special focus on ensemble techniques. Commonly used classifier ensembles combine outputs of base learning models in order to enhance the learning results. The multi-label classification problem introduces some new challenges to the ensemble learning methods. For instance, one needs to decide in which order is it better to aggregate the base learners - on a level of individual labels and then for the whole label sets, or the other way around. We discuss this issue and experimentally compare selected approaches. In the experiments, we use data from JRS'2012 Data Mining Competition, whose scope was topical classification of biomedical research papers, and as the base learners we utilize the models employed by the winners of this contest.
AB - In this paper, we investigate different approaches to multi-label classification of textual data, with a special focus on ensemble techniques. Commonly used classifier ensembles combine outputs of base learning models in order to enhance the learning results. The multi-label classification problem introduces some new challenges to the ensemble learning methods. For instance, one needs to decide in which order is it better to aggregate the base learners - on a level of individual labels and then for the whole label sets, or the other way around. We discuss this issue and experimentally compare selected approaches. In the experiments, we use data from JRS'2012 Data Mining Competition, whose scope was topical classification of biomedical research papers, and as the base learners we utilize the models employed by the winners of this contest.
KW - Data mining
KW - Ensemble learning
KW - Multi-label classification
KW - Topical classification
UR - http://www.scopus.com/inward/record.url?scp=84872734844&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84872734844&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-35527-1_26
DO - 10.1007/978-3-642-35527-1_26
M3 - Conference contribution
AN - SCOPUS:84872734844
SN - 9783642355264
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 306
EP - 317
BT - Advanced Data Mining and Applications - 8th International Conference, ADMA 2012, Proceedings
T2 - 8th International Conference on Advanced Data Mining and Applications, ADMA 2012
Y2 - 15 December 2012 through 18 December 2012
ER -