Modeli za predviđanje kraja zainteresiranosti igrača za digitalnu igru kroz analizu ponašanja

master's thesis

• Title (english): Models for predicting the end of players' interest for a digital game through behavioral analysis
• Author: Kuzma Mustač
• Mentor: Mirko Sužnjević (mentor)
• Committee member: Mirko Sužnjević (predsjednik povjerenstva)
• Committee member: Igor Sunday Pandžić (član povjerenstva)
• Committee member: Lea Skorin-Kapov (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2021-09-24, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

U ovom radu proveden je statistički pregled koristeći podatke jedne online mobilne igre. Statistički pregled podataka opravdao je vremenski i resursni trošak učenja i testiranja odabranih modela strojnog učenja. Cilj modela strojnog učenja je generalno predviđanje kraja zainteresiranosti igrača za igranjem i specifičniji slučaj predviđanja kraja zainteresiranosti igrača nakon potrošnje svih inicijalnih ulaznica unutar igre. Odabrani modeli strojnog učenja su: algoritam nasumične šume, algoritam... More K-najbližih susjeda i logistička regresija. Rezultati klasifikacije generalnog predviđanja kraja zainteresiranosti igrača pokazali su kako je algoritam nasumične šume sa vrijednošću AUC metrike od 0.77 i prosječnom preciznošću AP od 94% bio najbolji od odabranih modela i postigao zadovoljavajuće rezultate. Po rezultatima specifične analize kraja zainteresiranosti nakon potrošnje inicijalnih ulaznica kao najbolja istaknula se logistička regresija sa vrijednošću AUC metrike od 0.64 i prosječnom preciznošću od 53%. Less

Abstract (english)

In this paper, a statistical overview was initially made using the data gathered from an online mobile game. The statistical overview justified the time and resource costs associated with the training and testing of selected machine learning models whose aim is predicting the end of players’ interest in playing the game and a more specific case of predicting the end of player’s interest in playing the game after consuming all initially given in-game tickets. The selected machine learning... More models were: random forest algorithm, K-nearest neighbor algorithm and logistic regression. The results of the general end of players’ interest classification showed the random forest algorithm having the greatest AUC metric of 0.74 and the greatest average precision results of 94%. The specific case classification results highlighted the logistic regression having the highest AUC metric value of 0.64 and the highest average precision of 53%. Less

• Keywords: mobilne igre; strojno učenje; analiza ponašanja; klasifikacija
• Keywords (english): mobile games; machine learning; behavioral analysis; classification
• Language: croatian
• URN:NBN: urn:nbn:hr:168:761039
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-01-13 15:41:51