prikaz prve stranice dokumenta Estimacija osnovnog energetskog stanja molekula primjenom algoritama strojnog učenja
  Download
PDF 3.16 MB
master's thesis
Estimacija osnovnog energetskog stanja molekula primjenom algoritama strojnog učenja
2024. urn:nbn:hr:190:453075
Babić, Filip
University of Rijeka
Faculty of Engineering
Department of Automation and Electronics
Section of Electronics, Robotics and Automation

Cite this document

Babić, F. (2024). Estimacija osnovnog energetskog stanja molekula primjenom algoritama strojnog učenja (Master's thesis). Rijeka: University of Rijeka, Faculty of Engineering. Retrieved from https://urn.nsk.hr/urn:nbn:hr:190:453075

Babić, Filip. "Estimacija osnovnog energetskog stanja molekula primjenom algoritama strojnog učenja." Master's thesis, University of Rijeka, Faculty of Engineering, 2024. https://urn.nsk.hr/urn:nbn:hr:190:453075

Babić, Filip. "Estimacija osnovnog energetskog stanja molekula primjenom algoritama strojnog učenja." Master's thesis, University of Rijeka, Faculty of Engineering, 2024. https://urn.nsk.hr/urn:nbn:hr:190:453075

Babić, F. (2024). 'Estimacija osnovnog energetskog stanja molekula primjenom algoritama strojnog učenja', Master's thesis, University of Rijeka, Faculty of Engineering, accessed 12 March 2025, https://urn.nsk.hr/urn:nbn:hr:190:453075

Babić F. Estimacija osnovnog energetskog stanja molekula primjenom algoritama strojnog učenja [Master's thesis]. Rijeka: University of Rijeka, Faculty of Engineering; 2024 [cited 2025 March 12] Available at: https://urn.nsk.hr/urn:nbn:hr:190:453075

F. Babić, "Estimacija osnovnog energetskog stanja molekula primjenom algoritama strojnog učenja", Master's thesis, University of Rijeka, Faculty of Engineering, Rijeka, 2024. Available at: https://urn.nsk.hr/urn:nbn:hr:190:453075

Please login to the repository to save this object to your list.
Metadata
TitleEstimacija osnovnog energetskog stanja molekula primjenom algoritama strojnog učenja
Title (english)Estimation of ground state energy of molecules using machine learning algorithms
AuthorFilip Babić
MentorZlatan Car (mentor)
Committee memberSaša Vlahinić (predsjednik povjerenstva)
Committee memberZlatan Car (član povjerenstva)
Committee memberDario Matika (član povjerenstva)
GranterUniversity of Rijeka
Faculty of Engineering
(Department of Automation and Electronics)
(Section of Electronics, Robotics and Automation)
Rijeka
Defense date and country2024-03-20, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Electrical Engineering
Automation and Robotics
Abstract
Tema rada je estimacija osnovnog energetskog stanja molekula primjenom modela strojnog učenja. Napravljen je uvod u polje kvantne kemije i hipoteze su definirane te je odrađen pregled literature. Pregled se sastojao od uvoda i definicija iz kvantne mehanike kako bi se postavila temeljna razina znanja. Nakon toga je analiziran i obrađen set podataka dobiven korištenjem teorije funckionala gustoće. U toku rada nabrojane su metode strojnog učenja i opisan je generalni postupak kojim se takvi... More modeli izrađuju i treniraju. Obradom seta podataka korištenjem metoda redukcije dimenzija kao što su PCA sukcesivno su trenirani i optimizirani modeli estimacije te su u svakom tom koraku prezentirani rezultati. Kada se set podataka pročistio i modeli estimatora se optimizirali definirane su ansambl metode koje su se istrenirale u nekoliko kombinacija nakon čega su se prikazali rezultati i odabran je najuspješniji model. Metode su se evaluirale standardnim metrikama R2, MAE i RMSE i standardnom devijacijom istih. Odabirom najboljeg modela ansambla učitan je set podataka koji nije bio korišten za treniranje kako bi se potvrdilo uspješno istraživanje. U zadnjem poglavlju prokomentiran je postupak istraživanja i odluke iza određenih metoda. Less
Abstract (english)
The theme of this paper was estimating ground state energies of a molecule using machine learning methods. At the start of the paper the reader is introduced into the field of quantum chemistry and the hypothesis was defined followed by a brief look at the state of the art. Some ground rules were set in the way of theory and formulas to help in understanding the process. This was followed by an analysis and visualisation of the dataset that was generated on the basis of density... More functional theory. Methods of machine learning were described and the process of model selection and training was defined. After setting up the dataset methods of dimensionality reduction via PCA was explained and the methods were applied with succesive training and optimisation of the models followed by displaying the results. When the dataset was sufficienty reduced and estimators were well optimised ensemble methods were defined. These were trained and cross valdiated in a few different combination and the most efficient one was selected. The methods were evaluated by the standard evaluation metrics R2, MAE, RMSE and their standard deviation. After selecting the best ensemble method a fresh dataset was loaded and applied to the estimator to confirm a succesful research. Last chapter was dedicated to discussion and commentary on the process and reasoning behind certain decisions. Less
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:190:453075
Study programmeTitle: Graduate University Study of Electrical Engineering
Study programme type: university
Study level: graduate
Academic / professional title: magistar/magistra inženjer/inženjerka elektrotehnike (magistar/magistra inženjer/inženjerka elektrotehnike)
Type of resourceText
File originBorn digital
Access conditionsOpen access
Terms of useLicence Attribution 4.0 International (CC BY 4.0) icon
Created on2024-03-15 16:18:02
accessibility

closeAccessibilityrefresh

If you wish to permanently save changes, click on Save, if not - your settings will be reset when you restart the browser.