prikaz prve stranice dokumenta Grupiranje podataka
  Download
PDF 1.17 MB
master's thesis
Grupiranje podataka
2020. urn:nbn:hr:126:501750
Habijanić, Ana
Josip Juraj Strossmayer University of Osijek
Department of Mathematics
Chair of Applied Mathematics
Applied Mathematics Research Group

Cite this document

Habijanić, A. (2020). Grupiranje podataka (Master's thesis). Retrieved from https://urn.nsk.hr/urn:nbn:hr:126:501750

Habijanić, Ana. "Grupiranje podataka." Master's thesis, Josip Juraj Strossmayer University of Osijek, Department of Mathematics, 2020. https://urn.nsk.hr/urn:nbn:hr:126:501750

Habijanić, Ana. "Grupiranje podataka." Master's thesis, Josip Juraj Strossmayer University of Osijek, Department of Mathematics, 2020. https://urn.nsk.hr/urn:nbn:hr:126:501750

Habijanić, A. (2020). 'Grupiranje podataka', Master's thesis, Josip Juraj Strossmayer University of Osijek, Department of Mathematics, accessed 21 September 2022, https://urn.nsk.hr/urn:nbn:hr:126:501750

Habijanić A. Grupiranje podataka [Master's thesis]. Osijek: Josip Juraj Strossmayer University of Osijek, Department of Mathematics; 2020 [cited 2022 September 21] Available at: https://urn.nsk.hr/urn:nbn:hr:126:501750

A. Habijanić, "Grupiranje podataka", Master's thesis, Josip Juraj Strossmayer University of Osijek, Department of Mathematics, Osijek, 2020. Available at: https://urn.nsk.hr/urn:nbn:hr:126:501750

Please login to the repository to save this object to your list.
Metadata
TitleGrupiranje podataka
Title (english)Data Clustering
AuthorAna Habijanić
MentorDarija Marković (mentor)
Committee memberKristian Sabo (predsjednik povjerenstva)
Committee memberDarija Marković (član povjerenstva)
Committee memberSuzana Miodragović (član povjerenstva)
GranterJosip Juraj Strossmayer University of Osijek
Department of Mathematics
(Chair of Applied Mathematics)
(Applied Mathematics Research Group)
Osijek
Defense date and country2020-12-17, Croatia
Scientific / art field,
discipline and subdiscipline		NATURAL SCIENCES
Mathematics
Mathematical Logic and Accounting
Abstract
Cilj ovog rada je upoznavanje s tehnikama grupiranja podataka. Grupiranje podataka je proces u kojem od grupe različitih objekata stvaramo klastere sličnih objekata. Postoje različite vrste grupiranja, ovisno o podatcima s kojima radimo. Četiri glavne vrste su: particijsko i hijerarhijsko grupiranje te grupiranje temeljeno na gustoći ili mreži podataka. Particijsko grupiranje grupira podatke u klastere koje čine slični podatci. Najznačajniji algoritam ove skupine je algoritam K–srednjih vrijednosti. Nastavno na njega, nastale su i druge inače kao što su algoritmi K–medoida i neizrazitih C–srednjih vrijednosti. Hijerarhijsko grupiranje, s druge strane, stvara hijerarhiju klastera koja se prikazuje dendrogramom. Postoje dva pristupa: ”bottom-up” (aglomerativno) i ”top-down” (divizno) grupiranje koji odreduju kojim će se redoslijedom spajati, odnosno razdvajati klasteri. Kod klaster analize, potrebno je odrediti početne centroide te izabrati broj K za broj klastera. Za svaki od ta dva problema postoje razne metode koje nas dovode do optimalnog rješenja. Nadalje, grupiranja temeljena na gustoći, odnosno mreži podataka koriste informacije o gustoći, obliku te broju zadanih podataka. Uz njihove varijacije, najistaknutiji algoritmi su DBSCAN za gustoću podataka te STING algoritam za mrežu podataka.
Abstract (english)
The aim of this paper is to get acquainted with the data clustering techniques. clustering is the process of making a group of abstract objects into classes of similar objects. There are different types of clustering, depending on the type of data. The four main types are: partitional, hierarchical, density-based, and grid-based clustering. Partitional clustering clusters the dataset into clusters that contain similar data. The most significant algorithm is the K–means algorithm. Additional to this algorithm, other versions appeared, such as the K–medoids algorithm and fuzzy C–means algorithm. Hierarchical clustering, on the other hand, creates a hierarchy of clusters which is represented with a dendrogram. There are two approaches: the ”bottom-up” (agglomerative) and the ”top-down” (divisive) clustering which determines the order of joining or splitting the clusters. In the cluster analysis, we need to determine the initial centers and the number of clusters. For each of those problems, there are different methods that result in an optimal solution. Moreover, the density-based and the grid-based clustering are using information about the density, the shape, and the number of data. The most significant algorithms are the DBSCAN for the density-based and the STING for the grid-based clustering.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:126:501750
Study programmeTitle: Mathematics; specializations in: Financial and Statistical Mathematics, Mathematics and Computer Science, Industrial and Applied Mathematics
Course: Financial and Statistical Mathematics
Study programme type: university
Study level: graduate
Academic / professional title: magistar/magistra matematike (magistar/magistra matematike)
Type of resourceText
File originBorn digital
Access conditionsOpen access
Terms of useLicence In copyright icon
Created on2020-12-18 07:32:03