Analiza i predikcija aktivnosti korisnika sportske aplikacije

master's thesis

• Title (english): Analysis and Prediction of User Activity in Sports Application
• Author: Tomislav Buhiniček
• Mentor: Tomislav Burić (mentor)
• Committee member: Tomislav Burić (predsjednik povjerenstva)
• Committee member: Mervan Pašić (član povjerenstva)
• Committee member: Mate Puljiz (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2019-07-04, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Cilj ovog rada bio je predikcija kvalitete korisnika na temelju akcija unutar aplikacije u prvih 7 dana korištenja. Na početku su određene najvažnije akcije koje korisnik može napraviti u aplikaciji. Nakon što su skupljeni podaci, podaci su agregirani s obzirom na korisnika. Iz tako dobivenih podataka napravljena je analiza iz kojih država dolaze korisnici i kojim danima u tjednu. Nakon toga su rađeni modeli koji bi predvidjeli broj impresija na temelju broja akcija nakon 7 dana. Rezultati nisu bili dobri pa su korisnici podijeljeni u 5 kategorija ovisno o kvaliteti. Korištenjem više metoda (Multinomijalna logistička regresija, slučajne šume, Gradient Boosting i metoda glasanjem ), dobiveni su bolji rezultati.

Abstract (english)

The main goal of this master thesis was to predict the quality of users based on actions within the app in the first 7 days of use. Firstly, the most important actions that a user can perform in the app were defined. Once data were collected, they were grouped by the user and then aggregated. From these data it was analyzed from which country and on which days of the week the users utilized the app. After that, models were created that would predict the number of impressions based on the number of actions made by users after the first 7 days. The results were not good, so users were divided into 5 categories depending on their quality. Using multiple methods (Multinomial Logistic Regression, Random Forests, Gradient Boosting and Voting Classifier), better results were obtained.

• Keywords: korisnik; akcije; analiza; korelacija; modeli; predviđanje
• Keywords (english): a user; actions; analyzed; correlation; models; predicting
• Language: croatian
• URN:NBN: urn:nbn:hr:168:208063
• 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: 2020-01-07 17:07:36