prikaz prve stranice dokumenta Predviđanje korisničkih kupovnih navika zasnovano na metodama analize vremenskih serija
No public access
master's thesis
Predviđanje korisničkih kupovnih navika zasnovano na metodama analize vremenskih serija
Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing, 2019. urn:nbn:hr:168:414877
Kartelo, Josip
University of Zagreb
Faculty of Electrical Engineering and Computing

Institutional repository: FER Repository

Cite this document

Kartelo, J. (2019). Predviđanje korisničkih kupovnih navika zasnovano na metodama analize vremenskih serija (Master's thesis). Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing. Retrieved from https://urn.nsk.hr/urn:nbn:hr:168:414877

Kartelo, Josip. "Predviđanje korisničkih kupovnih navika zasnovano na metodama analize vremenskih serija." Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, 2019. https://urn.nsk.hr/urn:nbn:hr:168:414877

Kartelo, Josip. "Predviđanje korisničkih kupovnih navika zasnovano na metodama analize vremenskih serija." Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, 2019. https://urn.nsk.hr/urn:nbn:hr:168:414877

Kartelo, J. (2019). 'Predviđanje korisničkih kupovnih navika zasnovano na metodama analize vremenskih serija', Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, accessed 21 June 2024, https://urn.nsk.hr/urn:nbn:hr:168:414877

Kartelo J. Predviđanje korisničkih kupovnih navika zasnovano na metodama analize vremenskih serija [Master's thesis]. Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing; 2019 [cited 2024 June 21] Available at: https://urn.nsk.hr/urn:nbn:hr:168:414877

J. Kartelo, "Predviđanje korisničkih kupovnih navika zasnovano na metodama analize vremenskih serija", Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, Zagreb, 2019. Available at: https://urn.nsk.hr/urn:nbn:hr:168:414877

Metadata
TitlePredviđanje korisničkih kupovnih navika zasnovano na metodama analize vremenskih serija
Title (english)Prediction of Consumer Shopping Habits Based On Time Series Analysis
AuthorJosip Kartelo
MentorGoran Delač (mentor)
Committee memberGoran Delač (predsjednik povjerenstva)
Committee memberMarin Šilić (član povjerenstva)
Committee memberŽeljko Ilić (član povjerenstva)
GranterUniversity of Zagreb
Faculty of Electrical Engineering and Computing
Zagreb
Defense date and country2019-07-08, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Electrical Engineering
Abstract
U ovom radu napravljeno je predviđanje korisničkih kupovnih navika bazirano na metodama analize vremenske serije. Objašnjene su vremenske serije te je dan pregled potrebnih koraka za analizu vremenskih serija. Opisani su korišteni modeli predviđanja: metoda autoregresivnog intervalnog pomičnog prosjeka - ARIMA, metoda autoregresivnog intervalnog pomičnog prosjeka s vanjskim varijablama - SARIMAX, duboka neuronska mreža, model ekstremnog gradijentnog pojačanja - XGBOOST te Facebook Prophet. Pojašnjen je korišteni skup podataka. Detaljno je prikazan korak istraživanja i analize podataka te je vizualno popraćen grafičkim primjerima. Pokazan je proces modeliranja klasičnim tehnikama ARIMA i SARIMAX. Korišten je model duboke neuronske mreže iz biblioteke Tensorflow. Iz biblioteke XGBoost upotrijebljen je istoimeni model. Posljednji korišteni model je Facebook Prophet. Svi korišteni modeli pripadaju bibliotekama otvorenog koda. Korišteni modeli su detaljno opisani i vizualno popraćeni. Napravljena je usporedba rezultata korištenih metoda te je donesen zaključak.
Abstract (english)
In this thesis, the forecasting of customer shopping habits was done based on time series analysis methods. An overview of time series analysis theory was given with necessary phases included in building and deploying the forecasting model. The predictive models used for time series forecasting are: autoregressive interval moving average - ARIMA, seasonal autoregressive interval moving average with exogenous variables - SARIMAX, deep neural network, extreme gradient boosting - XGBoost and Facebook Prophet. The data sets used for predictions were clarified and explained. Exploratory data analysis is made and is shown in detail, visually accompanied by graphic examples. The modeling process is demonstrated by classical ARIMA and SARIMAX techniques. A deep neural network model from the Tensorflow library was used. Xgboost model was used from the XGBoost library. The last model used in this thesis is Facebook Prophet. All presented models are part of open source libraries. The models are described in detail with graphics and visualization. A methods' results comparison was made and the conclusion was given.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:168:414877
Study programmeTitle: Information and Communication Technology
Study programme type: university
Study level: graduate
Academic / professional title: magistar/magistra inženjer/inženjerka informacijske i komunikacijske tehnologije (magistar/magistra inženjer/inženjerka informacijske i komunikacijske tehnologije)
Type of resourceText
File originBorn digital
Access conditionsClosed access
Terms of useLicence In copyright icon
Public note
Created on2020-01-07 19:01:10