prikaz prve stranice dokumenta Usporedba metoda za grupiranje tekstualnih dokumenata
No public access
undergraduate thesis
Usporedba metoda za grupiranje tekstualnih dokumenata
Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing, 2017. urn:nbn:hr:168:179791
Bošnjak, Dario
University of Zagreb
Faculty of Electrical Engineering and Computing

Institutional repository: FER Repository

Cite this document

Bošnjak, D. (2017). Usporedba metoda za grupiranje tekstualnih dokumenata (Undergraduate thesis). Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing. Retrieved from https://urn.nsk.hr/urn:nbn:hr:168:179791

Bošnjak, Dario. "Usporedba metoda za grupiranje tekstualnih dokumenata." Undergraduate thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, 2017. https://urn.nsk.hr/urn:nbn:hr:168:179791

Bošnjak, Dario. "Usporedba metoda za grupiranje tekstualnih dokumenata." Undergraduate thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, 2017. https://urn.nsk.hr/urn:nbn:hr:168:179791

Bošnjak, D. (2017). 'Usporedba metoda za grupiranje tekstualnih dokumenata', Undergraduate thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, accessed 15 June 2024, https://urn.nsk.hr/urn:nbn:hr:168:179791

Bošnjak D. Usporedba metoda za grupiranje tekstualnih dokumenata [Undergraduate thesis]. Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing; 2017 [cited 2024 June 15] Available at: https://urn.nsk.hr/urn:nbn:hr:168:179791

D. Bošnjak, "Usporedba metoda za grupiranje tekstualnih dokumenata", Undergraduate thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, Zagreb, 2017. Available at: https://urn.nsk.hr/urn:nbn:hr:168:179791

Metadata
TitleUsporedba metoda za grupiranje tekstualnih dokumenata
Title (english)Comparison of Text Clustering Methods
AuthorDario Bošnjak
MentorKrešimir Pripužić (mentor)
Committee memberKrešimir Pripužić (predsjednik povjerenstva)
Committee memberDenis Salopek (član povjerenstva)
Committee memberMarko Pavelić (član povjerenstva)
GranterUniversity of Zagreb
Faculty of Electrical Engineering and Computing
Zagreb
Defense date and country2017-07-10, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Electrical Engineering
Abstract
Obrada i analiza dokumenata iz velikih kolekcija zahtijeva određenu organizaciju istih. Tomu služi grupiranje, postupak kojim će računalo, s obzirom na zadane parametre, odrediti grupu za dokumente iz kolekcije. Parametri se zadaju s obzirom na kolekciju i želje korisnika. Algoritam k-means jadan je od najpoznatijih algoritama, dok je poboljšanje dano algoritmom k-means++. Fuzzy k-means je meki algoritam grupiranja koji ovisno o zadanom parametru s određenim postotkom pripadnosti može svrstati dokumente u jednu ili više grupa. Algoritam multiple k-means++ svodi se na izvođenje algoritma k-means++ određeni broj puta. Navedeni algoritmi svrstavaju dokumente u zadani broj grupa, dok algoritam DBSCAN kao parametar ne prima željeni broj grupa, nego najmanji broj točaka u grupi i okolinu ɛ. Ako točka sadrži barem traženi broj točaka u svojoj okolini tada može postati jezgrom grupe, a u suprotnom će biti svrstana u neku grupu ili biti označena kao šum. Algoritam DBSCAN prilagodljiviji je od ostalih algoritama, ali je teško odrediti parametre s kojima bi se dobili dobri rezultati jer jako ovise o kolekciji.
Abstract (english)
Analysis of large collections can be a difficult task that can be simplified by dividing the collection into groups. Each group contains documents that are somehow similar. The process of dividing the collection can be done manually by a human, or automatically by a machine. Clustering is an automated process of dividing the collection into groups (clusters), with respect to parameters set by the user. K-means is a well known clustering algorithm, an improved version also exists, called k-means++. Both algorithms cluster a data point into one cluster. A fuzzy k-means algorithm is different, because it can cluster one data point into several clusters, depending on the user requirements. The DBSCAN algorithm, unlike aforementioned algorithms, does not require a fixed number of groups, but a minimum number of points in the group and ɛ neighborhood size. If a data point contains the minimum number of data points needed in it's neighborhood it can then become a cluster core, otherwise it will be grouped later in the process or labeled as noise. Both parameters are hard to set properly, because they severely depend on the collection.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:168:179791
Study programmeTitle: Computing
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 conditionsClosed access
Terms of useLicence In copyright icon
Public note
Created on2019-03-13 19:14:05