Učinkovitost učnih algoritmov na velikih podatkih

diplomsko delo

Jože Gobar (Author), Vili Podgorelec (Mentor), Sašo Karakatič (Co-mentor)

Abstract

Tematika te diplomske naloge so veliki podatki, karakteristike velikih podatkov in učni algoritmi, ki jih uporabljamo za klasifikacijo. V diplomski nalogi predstavljam tudi rezultate eksperimenta, s katerim sem ugotavljal učinkovitost učnih algoritmov na velike podatke. Učinkovitost algoritmov sem ovrednotil s klasifikacijsko točnostjo in časovnim izvajanjem učnih algoritmov na podatkovnih množicah. Iz pridobljenih rezultatov lahko sklepam, da se algoritmi glede na dane podatkovne množice različno obnašajo ter da je izbira učnega algoritma za analizo podatkovnih množic odvisna predvsem od problema in zastavljenega cilja.

Keywords

veliki podatki;karakteristike velikih podatkov;strojno učenje;klasifikacija;učinkovitost učnih algoritmov;

Data

Language:	Slovenian
Year of publishing:	2015
Typology:	2.11 - Undergraduate Thesis
Organization:	UM FERI - Faculty of Electrical Engineering and Computer Science
Publisher:	J. Gobar
UDC:	004.421:004.65(043.2)
COBISS:	19253270
Views:	866
Downloads:	66
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	EFFICIENCY OF LEARNING ALGORITHMS ON BIG DATA
Secondary abstract:	This thesis presents the subject of big data and its characteristics, learning algorithms applied in classification, as well as the results of an applied experiment in order to determine learning algorithms efficiency on big data. Algorithm efficiency has been assessed by classification accuracy and timely implementation of learning algorithms on datasets. The results indicate that algorithms perform differently considering given dataset and that preference of a learning algorithm intended for dataset analysis depends upon the posed problem and objective.
Secondary keywords:	big data characteristic;machine learning;classification;learning algorithms efficiency;
URN:	URN:SI:UM:
Type (COBISS):	Bachelor thesis/paper
Thesis comment:	Univ. v Mariboru, Fak. za elektrotehniko, računalništvo in informatiko, Informatika in informacijske tehnologije
Pages:	IX, 53 f.
ID:	9056978