Napredna pretraga

Pregled bibliografske jedinice broj: 849449

Employing different optimisation approaches for SMOTE parameter tuning


Zorić, Bruno; Bajer, Dražen; Martinović, Goran;
Employing different optimisation approaches for SMOTE parameter tuning // Proceedings of the 1st International Conference on Smart Systems and Technologies (SST) / Žagar, Drago ; Martinović, Goran ; Rimac Drlje, Snježana ; (ur.).
Osijek: Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2016. str. 191-196 (predavanje, međunarodna recenzija, cjeloviti rad (in extenso), znanstveni)


Naslov
Employing different optimisation approaches for SMOTE parameter tuning

Autori
Zorić, Bruno ; Bajer, Dražen ; Martinović, Goran ;

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
Proceedings of the 1st International Conference on Smart Systems and Technologies (SST) / Žagar, Drago ; Martinović, Goran ; Rimac Drlje, Snježana ; - Osijek : Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2016, 191-196

ISBN
978-1-5090-3718-6

Skup
International Conference on Smart Systems and Technologies (SST)

Mjesto i datum
Osijek, Hrvatska, 12-14.10.2016

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
SMOTE; bio-inspired optimisation algorithms; data imbalance; parameter tuning

Sažetak
When performing classification tasks on problems stemming from the real-world phenomena, issues requiring attention often arise due to the specific nature of the data. One common problem is the imbalance of datasets, the resolving of which is performed through several different approaches suggested in the literature. These approaches require the knowledge regarding data distribution and complexity to tune their parameters in order to better deal with the imbalance. This paper aims to give insight on employing various optimisation approaches in order to optimise parameters for the synthetic minority over-sampling technique. Several prominent optimisation approaches are evaluated on different real-world datasets in order to see whether the selection of a specific approach would produce considerably better results thus encouraging the favouring of one approach over the others.

Izvorni jezik
Engleski

Znanstvena područja
Računarstvo



POVEZANOST RADA


Ustanove
Fakultet elektrotehnike, računarstva i informacijskih tehnologija Osijek