prikaz prve stranice dokumenta USPOREDBA ALGORITAMA PREUZORKOVANJA ZASNOVANIH NA SMOTE ZA RJEŠAVANJE PROBLEMA NEURAVNOTEŽENOSTI KLASA
  Download
PDF 788.16 KB
undergraduate thesis
USPOREDBA ALGORITAMA PREUZORKOVANJA ZASNOVANIH NA SMOTE ZA RJEŠAVANJE PROBLEMA NEURAVNOTEŽENOSTI KLASA
Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2020. urn:nbn:hr:200:377270
Rojnić, Domagoj
Josip Juraj Strossmayer University of Osijek
Faculty of Electrical Engineering, Computer Science and Information Technology Osijek
Department of Software Engineering
Chair of Programming Languages and Systems

Institutional repository: Faculty of Electrical Engineering, Computer Science and Information Technology Osijek

Cite this document

Rojnić, D. (2020). USPOREDBA ALGORITAMA PREUZORKOVANJA ZASNOVANIH NA SMOTE ZA RJEŠAVANJE PROBLEMA NEURAVNOTEŽENOSTI KLASA (Undergraduate thesis). Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek. Retrieved from https://urn.nsk.hr/urn:nbn:hr:200:377270

Rojnić, Domagoj. "USPOREDBA ALGORITAMA PREUZORKOVANJA ZASNOVANIH NA SMOTE ZA RJEŠAVANJE PROBLEMA NEURAVNOTEŽENOSTI KLASA." Undergraduate thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2020. https://urn.nsk.hr/urn:nbn:hr:200:377270

Rojnić, Domagoj. "USPOREDBA ALGORITAMA PREUZORKOVANJA ZASNOVANIH NA SMOTE ZA RJEŠAVANJE PROBLEMA NEURAVNOTEŽENOSTI KLASA." Undergraduate thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2020. https://urn.nsk.hr/urn:nbn:hr:200:377270

Rojnić, D. (2020). 'USPOREDBA ALGORITAMA PREUZORKOVANJA ZASNOVANIH NA SMOTE ZA RJEŠAVANJE PROBLEMA NEURAVNOTEŽENOSTI KLASA', Undergraduate thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, accessed 28 April 2024, https://urn.nsk.hr/urn:nbn:hr:200:377270

Rojnić D. USPOREDBA ALGORITAMA PREUZORKOVANJA ZASNOVANIH NA SMOTE ZA RJEŠAVANJE PROBLEMA NEURAVNOTEŽENOSTI KLASA [Undergraduate thesis]. Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek; 2020 [cited 2024 April 28] Available at: https://urn.nsk.hr/urn:nbn:hr:200:377270

D. Rojnić, "USPOREDBA ALGORITAMA PREUZORKOVANJA ZASNOVANIH NA SMOTE ZA RJEŠAVANJE PROBLEMA NEURAVNOTEŽENOSTI KLASA", Undergraduate thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, Osijek, 2020. Available at: https://urn.nsk.hr/urn:nbn:hr:200:377270

Metadata
TitleUSPOREDBA ALGORITAMA PREUZORKOVANJA ZASNOVANIH NA SMOTE ZA RJEŠAVANJE PROBLEMA NEURAVNOTEŽENOSTI KLASA
Title (english)Comparison of oversampling algorithms based on SMOTE for dealing with the class imbalance problem.
AuthorDomagoj Rojnić
MentorZdravko Krpić (mentor)
GranterJosip Juraj Strossmayer University of Osijek
Faculty of Electrical Engineering, Computer Science and Information Technology Osijek
(Department of Software Engineering)
(Chair of Programming Languages and Systems)
Osijek
Defense date and country2020-10-05, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Data Processing
Abstract
U radu je napravljena usporedba algoritama preuzorkovanja zasnovanih na SMOTE za rješavanje problema neuravnoteženosti klasa. Rad daje detaljan opis pojma klasifikacije podataka, problema neuravnoteženosti klasa, algoritma k-najbližih susjeda kao i algoritma SMOTE i njegovih inačica, Borderline-SMOTE i Safe-level-SMOTE. Također su opisane različite mjere učinkovitosti uspoređivanih algoritama. Za potrebe rada ostvareno je programsko rješenje pomoću kojeg se određuje prosječna F-mjera primjene algoritma preuzorkovanja nad određenim stvarnim skupom podataka. Programsko rješenje ostvareno je u .NET okruženju koristeći programski jezik C#. Usporedba četiri algoritma preuzorkovanja uz klasifikaciju bez primjene preuzorkovanja provedena je nad šest skupova podataka različite brojnosti i omjera neuravnoteženosti. Rezultatima je dana prosječna F-mjera iz koje je vidljivo održavanje konzistentnosti rezultata neovisno o algoritmu prilikom čega je SMOTE jedini postigao maksimalnu F-mjeru u dva slučaja, dok njegove inačice nisu pokazale znatno bolje niti znatno lošije rezultate u odnosu na originalni SMOTE.
Abstract (english)
This paper provides comparison of SMOTE-based oversampling algorithms for solving class imbalance problem. Paper gives detail description of terms data classification, class imbalance problem, k-nearest neighbors algorithm and SMOTE algorithm with its improved versions, Borderline-SMOTE and Safe-level-SMOTE. For the purpose of this paper, software was implemented to calculate average F-measure of oversampling algorithm over real-world datasets. Software was implemented in .NET framework using C# programming lanuage. Comparison of four oversampling algorithms added with method of no oversampling was conducted over six datasets with different number of samples and different imbalance ratios. Results provide average F-measure, which shows maintaned consistency of the results independent of the algorithm, whereby only SMOTE achieved maximum F-measure in two cases while its improvements did not show significantly better or worse results compared to the original SMOTE.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:200:377270
Study programmeTitle: Undergraduate University Study Programme in Computer Engineering
Study programme type: university
Study level: undergraduate
Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva)
Type of resourceText
File originBorn digital
Access conditionsOpen access
Terms of useLicence In copyright icon
Created on2020-11-16 16:25:15