prikaz prve stranice dokumenta Učinkovitost tehnika predobrade podataka u izgradnji modela predviđanja
  Download
PDF 1009.33 KB
undergraduate thesis
Učinkovitost tehnika predobrade podataka u izgradnji modela predviđanja
Rijeka: University of Rijeka, Faculty of Engineering, 2023. urn:nbn:hr:190:339915
Višković, Kristina
University of Rijeka
Faculty of Engineering

Institutional repository: Repository of the University of Rijeka, Faculty of Engineering

Cite this document

Višković, K. (2023). Učinkovitost tehnika predobrade podataka u izgradnji modela predviđanja (Undergraduate thesis). Rijeka: University of Rijeka, Faculty of Engineering. Retrieved from https://urn.nsk.hr/urn:nbn:hr:190:339915

Višković, Kristina. "Učinkovitost tehnika predobrade podataka u izgradnji modela predviđanja." Undergraduate thesis, University of Rijeka, Faculty of Engineering, 2023. https://urn.nsk.hr/urn:nbn:hr:190:339915

Višković, Kristina. "Učinkovitost tehnika predobrade podataka u izgradnji modela predviđanja." Undergraduate thesis, University of Rijeka, Faculty of Engineering, 2023. https://urn.nsk.hr/urn:nbn:hr:190:339915

Višković, K. (2023). 'Učinkovitost tehnika predobrade podataka u izgradnji modela predviđanja', Undergraduate thesis, University of Rijeka, Faculty of Engineering, accessed 23 November 2024, https://urn.nsk.hr/urn:nbn:hr:190:339915

Višković K. Učinkovitost tehnika predobrade podataka u izgradnji modela predviđanja [Undergraduate thesis]. Rijeka: University of Rijeka, Faculty of Engineering; 2023 [cited 2024 November 23] Available at: https://urn.nsk.hr/urn:nbn:hr:190:339915

K. Višković, "Učinkovitost tehnika predobrade podataka u izgradnji modela predviđanja", Undergraduate thesis, University of Rijeka, Faculty of Engineering, Rijeka, 2023. Available at: https://urn.nsk.hr/urn:nbn:hr:190:339915

Metadata
TitleUčinkovitost tehnika predobrade podataka u izgradnji modela predviđanja
Title (english)Efficiency of data-preprocessing techniques in building a predictive model
AuthorKristina Višković
MentorGoran Mauša (mentor)
Committee memberMiroslav Joler (predsjednik povjerenstva)
Committee memberGoran Mauša (član povjerenstva)
GranterUniversity of Rijeka
Faculty of Engineering
Rijeka
Defense date and country2023, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Abstract
Zadatak završnog rada bio je ispitati učinkovitost tehnika predobrade podataka u izgradnji modela predviđanja. Proces dubinske analize podataka objašnjen je procesnim modelom CRISP-DM. Dodatno, obrađena su načela održive dubinske analize podataka vezana uz ključne faze procesa dubinske analize podataka. Proučeni su postupci predobrade podataka u dubinskoj analizi podataka, točnije tehnike čišćenja podataka kojima se uklanjaju instance i značajke. Nad karakterističnim skupovima podataka primjenjene su tehnike uklanjanja instanci i značajki koje sadrže nedostajuće vrijednosti, uklanjanje dupliciranih instanci i tehnika odabira najboljih značajki. Korišteni su algoritmi strojnog učenja stablo odluke i nasumična šuma te algoritam podizanja gradijenata temeljen na histogramu (engl. Histogram-based Gradient Boosting). Izmjeren je utrošak energije, memorije i vremena za scenarije izvornih i očišćenih podataka. Zaključeno je da je kvalitetnom predobradom podataka moguće smanjiti utrošak energije i vremena bez ugrožavanja performansi modela predviđanja.
Abstract (english)
The task of this thesis was to examine the efficiency of data-preprocessing techniques in building a prediction model. Data mining is explained using the CRISP-DM process model. The thesis also covers the principles of green data mining related to the key stages of the data mining process. Data preprocessing techniques, specifically data cleaning techniques used to remove instances and features (instance and feature selection techniques), were researched. The techniques of removing instances and features that contain missing values, removing duplicate instances and feature selection were applied to the characteristic data sets. Decision tree, random forest and histogram-based gradient boosting machine learning algorithms were used. The consumption of energy, memory and time was measured for the original and cleaned data scenarios. It was concluded that with high-quality data preprocessing, it is possible to reduce the consumption of energy and time without disrupting the performance of the prediction model.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:190:339915
Study programmeTitle: Computer Engineering
Study programme type: university
Study level: undergraduate
Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva)
Type of resourceText
File originBorn digital
Access conditionsOpen access
Terms of useLicence Attribution 4.0 International (CC BY 4.0) icon
Created on2023-09-12 15:24:03