Zasnova in razvoj sistema za rudarjenje mnenja, s pomočjo besednih vektorjev in nevronskih mrež

magistrsko delo

Miha Hozjan (Author), Borko Bošković (Mentor), Janez Brest (Co-mentor)

Abstract

V zadnjih letih je, predvsem s porastom socialnih medijev, analiza sentimenta postala ena izmed glavnih vej obdelave naravnega jezika. Intenzivno se uporablja na različnih področjih, med drugim tudi v političnem prostoru. V sklopu magistrske naloge smo izdelali sistem za rudarjenje mnenja, ki uspešno razvrsti komentarje, zapisane v slovenskem jeziku, v tri kategorije, in sicer pozitivne, negativne in nevtralne. Po proučitvi strokovne in znanstvene literature ter razvoju in primerjavi različnih modelov nam je uspelo pokazati, da lahko z uvedbo besednih vektorjev v kombinaciji z nevronskimi mrežami občutno izboljšamo delovanje takšnega sistema.

Keywords

analiza sentimenta;besedni vektorji;nevronske mreže;obdelava naravnega jezika;magistrske naloge;

Data

Language:	Slovenian
Year of publishing:	2020
Typology:	2.09 - Master's Thesis
Organization:	UM FERI - Faculty of Electrical Engineering and Computer Science
Publisher:	[M. Hozjan]
UDC:	004.8:004.93(043.2)
COBISS:	37253891
Views:	356
Downloads:	55
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	Design and development of an opinion mining system with the use of word vectors and neural networks
Secondary abstract:	Sentiment analysis has become one of the most important branches of natural language processing, especially with the rise of social media over the past few years. It is extensively used in various fields, including politics. As part of the master's thesis, we have developed an opinion mining system that successfully classifies comments written in the Slovenian language into three different categories, namely positive, negative and neutral. After reviewing the technical and scientific literature along with the implementation and comparison of different models, we have been able to demonstrate that the introduction of word vectors in combination with neural networks can significantly improve the performance of such a system.
Secondary keywords:	sentiment analysis;word vectors;neural networks;natural language processing;
Type (COBISS):	Master's thesis/paper
Thesis comment:	Univ. v Mariboru, Fak. za elektrotehniko, računalništvo in informatiko, Računalništvo in informacijske tehnologije
Pages:	VIII, 47 f.
ID:	11945911