prikaz prve stranice dokumenta Radni okvir za umjetne neuronske mreže
No public access
undergraduate thesis
Radni okvir za umjetne neuronske mreže
Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing, 2020. urn:nbn:hr:168:676098
Marčinković, David
University of Zagreb
Faculty of Electrical Engineering and Computing

Institutional repository: FER Repository

Cite this document

Marčinković, D. (2020). Radni okvir za umjetne neuronske mreže (Undergraduate thesis). Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing. Retrieved from https://urn.nsk.hr/urn:nbn:hr:168:676098

Marčinković, David. "Radni okvir za umjetne neuronske mreže." Undergraduate thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, 2020. https://urn.nsk.hr/urn:nbn:hr:168:676098

Marčinković, David. "Radni okvir za umjetne neuronske mreže." Undergraduate thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, 2020. https://urn.nsk.hr/urn:nbn:hr:168:676098

Marčinković, D. (2020). 'Radni okvir za umjetne neuronske mreže', Undergraduate thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, accessed 26 December 2024, https://urn.nsk.hr/urn:nbn:hr:168:676098

Marčinković D. Radni okvir za umjetne neuronske mreže [Undergraduate thesis]. Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing; 2020 [cited 2024 December 26] Available at: https://urn.nsk.hr/urn:nbn:hr:168:676098

D. Marčinković, "Radni okvir za umjetne neuronske mreže", Undergraduate thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, Zagreb, 2020. Available at: https://urn.nsk.hr/urn:nbn:hr:168:676098

Metadata
TitleRadni okvir za umjetne neuronske mreže
Title (english)Framework for Artificial Neural Networks
AuthorDavid Marčinković
MentorZoran Kalafatić (mentor)
Committee memberZoran Kalafatić (predsjednik povjerenstva)
Committee memberJosip Šarić (član povjerenstva)
Committee memberIvan Grubišić (član povjerenstva)
GranterUniversity of Zagreb
Faculty of Electrical Engineering and Computing
Zagreb
Defense date and country2020-07-06, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Abstract
Umjetne neuronske mreže su računalni sustavi temeljeni na matematičkim modelima koji su inspirirani biološkim neuronskim mrežama, a svoju primjenu nalaze u područjima umjetne inteligencije, obrade podataka, robotici te mnogim drugim područjima znanosti. Sastoje od jednog ulaznog i izlaznog sloja te jednog ili više skrivenih slojeva. Broj neurona ulaznog sloja određen je veličinom ulaznih podataka, dok je broj neurona na izlazu jednak broju razreda koje želimo odrediti. Broj skrivenih slojeva i broj neurona u pojedinom skrivenom sloju je specifičan za svaki problem i jedan je od parametara kojeg treba pažljivo odrediti. Težine poveznica između pojedinih slojeva na samom početku inicijaliziramo nasumično s pomno odabranom funkcijom razdiobe. Koristimo aktivacijsku funkciju kako bismo izlaz iz neuronske mreže preslikali na željeni interval. Stopa učenja je jedan od najvažnijih parametara kada govorimo o neuronskim mrežama. O stopi učenja ovisi brzina procesa učenja i konačna preciznost neuronske mreže. Neuronske mreže rade na principu nadziranog učenja, što se odnosi na činjenicu da mreži u procesu učenja dajemo označene podatke. Neuronske mreže daju jako dobre rezultate u procesu raspoznavanja objekata na fotografija.
Abstract (english)
Artificial neural networks are computer systems based on mathematical models which are inspired by biological neural networks, and find their application in the fields of artificial intelligence, data processing, robotics and many other fields of science. They consist of one input and one output layer, and one or more hidden layers. The number of input layer neurons is determined by the size of the input data, while the number of output neurons is equal to the number of classes we want to decide. The number of hidden layers and the number of neurons in each hidden layer is problem specific, and is one of the most important parameters that has to be carefully chosen. Weights between layers are initialized randomly using some random distribution function. Activation function is used to map neural network output to some desired range. Learning rate is one of the most important parameters related to neural networks. Learning rate determines the speed of learning process, as well as final prediction ability. Neural networks work on the principle of supervised learning, which refers to the fact that in the learning process we give the network labeled data . Neural network give very good results in the process of object recognition and classification.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:168:676098
Study programmeTitle: Computing
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 conditionsClosed access
Terms of useLicence In copyright icon
Public note
Created on2023-01-14 09:23:03