prikaz prve stranice dokumenta Primjena neuronskih mreža za predviđanje duljine studija studenata preddiplomskog studija
Access restricted to students and staff of home institution
master's thesis
Primjena neuronskih mreža za predviđanje duljine studija studenata preddiplomskog studija
2023. urn:nbn:hr:124:969000
Cvijetić, Nikolina
University of Split
Faculty of economics Split

Cite this document

Cvijetić, N. (2023). Primjena neuronskih mreža za predviđanje duljine studija studenata preddiplomskog studija (Master's thesis). Split: University of Split, Faculty of economics Split. Retrieved from https://urn.nsk.hr/urn:nbn:hr:124:969000

Cvijetić, Nikolina. "Primjena neuronskih mreža za predviđanje duljine studija studenata preddiplomskog studija." Master's thesis, University of Split, Faculty of economics Split, 2023. https://urn.nsk.hr/urn:nbn:hr:124:969000

Cvijetić, Nikolina. "Primjena neuronskih mreža za predviđanje duljine studija studenata preddiplomskog studija." Master's thesis, University of Split, Faculty of economics Split, 2023. https://urn.nsk.hr/urn:nbn:hr:124:969000

Cvijetić, N. (2023). 'Primjena neuronskih mreža za predviđanje duljine studija studenata preddiplomskog studija', Master's thesis, University of Split, Faculty of economics Split, accessed 24 December 2024, https://urn.nsk.hr/urn:nbn:hr:124:969000

Cvijetić N. Primjena neuronskih mreža za predviđanje duljine studija studenata preddiplomskog studija [Master's thesis]. Split: University of Split, Faculty of economics Split; 2023 [cited 2024 December 24] Available at: https://urn.nsk.hr/urn:nbn:hr:124:969000

N. Cvijetić, "Primjena neuronskih mreža za predviđanje duljine studija studenata preddiplomskog studija", Master's thesis, University of Split, Faculty of economics Split, Split, 2023. Available at: https://urn.nsk.hr/urn:nbn:hr:124:969000

Metadata
TitlePrimjena neuronskih mreža za predviđanje duljine studija studenata preddiplomskog studija
Title (english)Application of neural networks to predict the length of studies of undergraduate students
AuthorNikolina Cvijetić
MentorMarko Hell (mentor)
Committee memberTea Šestanović (član povjerenstva)
Committee memberMaja Mihaljević Kosor (član povjerenstva)
GranterUniversity of Split
Faculty of economics Split
Split
Defense date and country2023, Croatia
Scientific / art field,
discipline and subdiscipline		SOCIAL SCIENCES
Economics
Abstract
Neuronske mreže postaju sve moćniji alat dubokog strojnog učenja, sve više dobivaju na važnosti u različitim područjima znanosti pružajući razne mogućnosti za rješavanje kompleksnih problema. Jedna od primjena je u visokom obrazovanju. Cilj ovoga rada je bio ispitati da li je moguće uz dostupne podatke izgraditi neuronsku mrežu koja će uspješno predviđati trajanje studija svakog pojedinog studenta na temelju rezultata s prve godine preddiplomskog studija. U ovom istraživanju razvijena su dva modela s različitim ulaznim parametrima, oba su klasifikacijska modela. Kako bi se modeli usporedili koristila se matrica uspješnosti te metrike kao što su preciznost i opoziv. Temeljem analize i usporedbe odabran je model 1 jer je model 2 pokazivao znakove prenaučenosti. Rezultati ovog rada pokazuju da je moguće izgraditi neuronsku mrežu koja može predviđati hoće li studenti završiti unutar predviđenog roka, van roka ili da neće završiti preddiplomski studij u solidnu točnost. Također, daljnji rad i razvoj modela uz proširenje ulaznih parametara moglo bi dovesti do izgradnje još uspješnijeg modela.
Abstract (english)
Neural networks are becoming an increasingly powerful tool in deep machine learning, gaining significance in various scientific domains by providing diverse possibilities for solving complex problems. They are also used in higher education. The objective of this study was to examine whether it is possible to construct a neural network that successfully predicts the duration of each individual student's study based on their performance in the first year of undergraduate studies. Two models with different input parameters were developed for this research, both of which were classification models. To compare the models, a success matrix and metrics such as precision and recall were utilized. Based on the analysis and comparison, model 1 was selected as model 2 exhibited signs of overfitting. Results of this study demonstrate that it is possible to build a solid neural network capable of predicting whether students will complete their undergraduate studies within the projected timeframe, beyond it, or not at all. Furthermore, further work and model development with expanded input parameters could lead to the creation of an even more effective model.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:124:969000
Study programmeTitle: Business Studies
Study programme type: university
Study level: graduate
Academic / professional title: magistar/magistra ekonomije (magistar/magistra ekonomije)
Type of resourceText
File originBorn digital
Access conditionsAccess restricted to students and staff of home institution
Terms of useLicence In copyright icon
Created on2023-09-22 13:39:42