prikaz prve stranice dokumenta PRIMJENA STROJNOG UČENJA U SVRHU DETEKTIRANJA ANOMALIJA U STREAMING PODACIMA
Access restricted to students and staff of home institution
master's thesis
PRIMJENA STROJNOG UČENJA U SVRHU DETEKTIRANJA ANOMALIJA U STREAMING PODACIMA
2017. urn:nbn:hr:145:279368
Majić, Stefani
Josip Juraj Strossmayer University of Osijek
Faculty of Economics and Business in Osijek
Chair of Quantitative Methods and Informatics

Cite this document

Majić, S. (2017). PRIMJENA STROJNOG UČENJA U SVRHU DETEKTIRANJA ANOMALIJA U STREAMING PODACIMA (Master's thesis). Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Economics and Business in Osijek. Retrieved from https://urn.nsk.hr/urn:nbn:hr:145:279368

Majić, Stefani. "PRIMJENA STROJNOG UČENJA U SVRHU DETEKTIRANJA ANOMALIJA U STREAMING PODACIMA." Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Economics and Business in Osijek, 2017. https://urn.nsk.hr/urn:nbn:hr:145:279368

Majić, Stefani. "PRIMJENA STROJNOG UČENJA U SVRHU DETEKTIRANJA ANOMALIJA U STREAMING PODACIMA." Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Economics and Business in Osijek, 2017. https://urn.nsk.hr/urn:nbn:hr:145:279368

Majić, S. (2017). 'PRIMJENA STROJNOG UČENJA U SVRHU DETEKTIRANJA ANOMALIJA U STREAMING PODACIMA', Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Economics and Business in Osijek, accessed 14 November 2024, https://urn.nsk.hr/urn:nbn:hr:145:279368

Majić S. PRIMJENA STROJNOG UČENJA U SVRHU DETEKTIRANJA ANOMALIJA U STREAMING PODACIMA [Master's thesis]. Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Economics and Business in Osijek; 2017 [cited 2024 November 14] Available at: https://urn.nsk.hr/urn:nbn:hr:145:279368

S. Majić, "PRIMJENA STROJNOG UČENJA U SVRHU DETEKTIRANJA ANOMALIJA U STREAMING PODACIMA", Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Economics and Business in Osijek, Osijek, 2017. Available at: https://urn.nsk.hr/urn:nbn:hr:145:279368

Please login to the repository to save this object to your list.
Metadata
TitlePRIMJENA STROJNOG UČENJA U SVRHU DETEKTIRANJA ANOMALIJA U STREAMING PODACIMA
Title (english)USING MACHINE LEARNING FOR ANOMALY DETECTION IN STREAMING DATA
AuthorStefani Majić
MentorMarijana Zekić Sušac (mentor)
Committee memberMarijana Zekić Sušac (predsjednik povjerenstva)
Committee memberSanja Pfeifer (član povjerenstva)
Committee memberDomagoj Sajter (član povjerenstva)
GranterJosip Juraj Strossmayer University of Osijek
Faculty of Economics and Business in Osijek
(Chair of Quantitative Methods and Informatics)
Osijek
Defense date and country2017-09-11, Croatia
Scientific / art field,
discipline and subdiscipline		SOCIAL SCIENCES
Economics
Business Informatics
Abstract
Cilj ovoga rada je istraživanje upotrebe algoritama strojnog učenja za detektiranje anomalija u streaming podacima. Streaming podaci su slijed stavki koje su u stvarnom vremenu (eng. real-time), kontinuirani te poredani implicitno po vremenu dolasku ili eksplicitno po vremenskoj oznaci (eng. timestamp). Aktualnost ovog rada pronalazi se u sve većem porastu količine dostupnih podataka porastom upotrebe big data i interneta stvari. U radu su predstavljena istraživanja brojnih autora koji su istraživali ovu tematiku te algoritmi, alati i metode kao i problemi implementacije algoritama strojnog učenja na streaming podacima. U praktičnom dijelu rada obrađene su metoda potpornih vektora i analiza glavnih komponenti (PCA) kako bi se identificirala njihova uspješnost u radu sa streaming podacima. Metoda potpornih vektora ima brojnih prednosti, ali i nedostataka, no ono što ju čini zanimljivom je njezina primjenjivost na skupu podataka u kojemu nema mnogo anomalija jer u stvarnosti nema puno sustava koji bi si dopustili velik broj abnormalnih podataka i to najčešće zbog troškova koje bi to prouzrokovalo. Kreirani model prepoznaje anomalije sa ukupnom stopom točnosti od 96,4% na uzorku za validaciju. Iako je područje u posljednje vrijeme dosta istraživano ima još dosta mjesta za razvoj, a uzevši u obzir eksponencijalni rast količine podataka, za pretpostaviti je kako će se o ovoj temi još dosta raspravljati.
Abstract (english)
The aim of this thesis is to explore the use of machine leaning algorithms for detecting anomalies in streaming data. Streaming data is a real-time, continuous sequence of items that are ordered implicitly by arrival time or explicitly by timestamp. The actuality of this thesis lies in an ever-increasing amount of available data because of the increase of using Big Data and Internet of Things. The paper presents research by numerous authors who have been researching this topic, algorithms, tools and methods as well as the problems of deploying and implementing machine learning algorithms on streaming data. In the practical part of the paper, the Support Vector Machine and Principal Component Analysis were used to identify their performance in streaming data. The Support Vector Machines have many advantages and disadvantages, but what makes it interesting is that it can be applied to a set of data in which there is not much data classified as anomalies, since in real life no system would allow a large number of abnormal data due to security and high costs. The created model shows an accuracy of 96,4% on validation data. Although this area has been explored for quite some time, there is still plenty of room for development if the exponential increase in data volume is considered. Therefore, it can be assumed that this topic is still to be discussed.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:145:279368
Study programmeTitle: Business economy; specializations in: Business Informatics
Course: Business Informatics
Study programme type: university
Study level: graduate
Academic / professional title: Magistar/magistra ekonomije (Magistar/magistra ekonomije)
Type of resourceText
File originBorn digital
Access conditionsAccess restricted to students and staff of home institution
Terms of useLicence In copyright icon
Created on2017-10-11 11:29:36