prikaz prve stranice dokumenta Postupak optimiziranog izbora oglasa na temelju stabala odlučivanja
  Download
PDF 1.77 MB
master's thesis
Postupak optimiziranog izbora oglasa na temelju stabala odlučivanja
2017. urn:nbn:hr:200:633437
Babić, Dominik
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

Cite this document

Babić, D. (2017). Postupak optimiziranog izbora oglasa na temelju stabala odlučivanja (Master's thesis). Retrieved from https://urn.nsk.hr/urn:nbn:hr:200:633437

Babić, Dominik. "Postupak optimiziranog izbora oglasa na temelju stabala odlučivanja." Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2017. https://urn.nsk.hr/urn:nbn:hr:200:633437

Babić, Dominik. "Postupak optimiziranog izbora oglasa na temelju stabala odlučivanja." Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2017. https://urn.nsk.hr/urn:nbn:hr:200:633437

Babić, D. (2017). 'Postupak optimiziranog izbora oglasa na temelju stabala odlučivanja', Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, accessed 18 March 2023, https://urn.nsk.hr/urn:nbn:hr:200:633437

Babić D. Postupak optimiziranog izbora oglasa na temelju stabala odlučivanja [Master's thesis]. Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek; 2017 [cited 2023 March 18] Available at: https://urn.nsk.hr/urn:nbn:hr:200:633437

D. Babić, "Postupak optimiziranog izbora oglasa na temelju stabala odlučivanja", Master's thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, Osijek, 2017. Available at: https://urn.nsk.hr/urn:nbn:hr:200:633437

Please login to the repository to save this object to your list.
Metadata
TitlePostupak optimiziranog izbora oglasa na temelju stabala odlučivanja
Title (english)Optimized selection of ads based on decision trees
AuthorDominik Babić
MentorZdravko Krpić (mentor)
Committee memberZdravko Krpić (predsjednik povjerenstva)
Committee memberMirko Kohler (član povjerenstva)
Committee memberBruno Zorić (član povjerenstva)
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 country2017-05-26, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Program Engineering
Abstract
Na području Internet oglašavanja često se javlja problem pronalaska efektivne i efikasne metode dostave oglasa potencijalnim korisnicima. Uobičajeno rješenje tog problema je primjena specijaliziranih sustava za oglašavanje kojima se nastoji dostavljati oglase na temelju prikupljenih informacija o korisniku. Cilj ovoga rada je primjenom stabala odlučivanja, iterativnih algoritama strojnog učenja, realizirati sustav koji na temelju prikupljenih informacija o prijašnjim korisnicima optimizira prikazivanje oglasa budućim korisnicima. Različitim pristupima treniranju stabla odlučivanja prikupljeni su rezultati za operatore Decision Tree, Decision Stump i Random Forest samostalno te u kombinaciji s ansambl operatorima AdaBoost i Bagging. Također su dani rezultati za Vote, Stacking i Gradient Boosted Trees ansambl operatore te za Forward Selection operator koji daje bolji uvid u korisnost pojedinih atributa pri treniranju stabala odlučivanja. Analizom rezultata i prikupljenih podataka predlažu se metode kojima bi se povećala točnost sustava temeljenog na stablima odlučivanja.
Abstract (english)
In the field of Internet advertising often emerges a problem of finding effective and efficient methods of ad delivery to potential customers. Typically the solution to this problem is the use of specialized systems for advertising where ad delivery is based on the information collected about the user. The objective of this paper is to create a system by using decision trees and iterative machine learning algorithms that, based on the information gathered about previous customers, optimizes ad delivery to the future customers. Different approaches for training the decision trees are used and results are collected for Decision Tree, Decision Stump and Random Forest operators alone and in their combination with the ensemble operators such as AdaBoost and Bagging. The results are also provided for the Vote, Stacking and Gradient Boosted Trees ensemble operators and the Forward Selection operator which provides better insight into the usefulness of certain attributes during the training of decision trees. By analyzing the results and the collected data, methods are proposed which would increase the accuracy of a system based on decision trees.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:200:633437
Study programmeTitle: Graduate University Study Programme in Computer Engineering
Study programme type: university
Study level: graduate
Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
Type of resourceText
File originBorn digital
Access conditionsOpen access
Terms of useLicence In copyright icon
Created on2017-06-20 15:04:32