Modeli obrade prirodnog jezika za detekciju sarkazma na društvenim mrežama

undergraduate thesis

• Title (english): Natural Language Processing Models for Sarcasm Detection on Social Media
• Author: Dunja Šmigovec
• Mentor: Jan Šnajder (mentor)
• Committee member: Jan Šnajder (predsjednik povjerenstva)
• Committee member: Ivan Crnomarković (član povjerenstva)
• Committee member: Zoran Medić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2021-07-05, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Obrada prirodnog jezika je grana umjetne inteligencije, koja pomaže računalima u određivanju sentimenta te značajnih dijelova jezika. Detekcija sarkazma specifičan je slučaj analize sentimenta i predstavlja izazovan zadatak u obradi prirodnog jezika. Budući da sarkazam koristi suprotni sentiment, teško ga je razumjeti, ne samo za strojeve, već i za ljude. U ovom radu ispitane su metode strojnog i dubokog učenja za klasifikaciju sarkazma u objavama na društvenim mrežama. Naša dva glavna pristupa uključuju BiLSTM modele, koji uzimaju u obzir odgovor i različite duljine konteksta. Zaključujemo da previše konteksta smanjuje točnost modela. Primjećujemo da pristupi dubokog učenja daju bolje rezultate od pristupa strojnog učenja.

Abstract (english)

Natural language processing is a branch of artificial intelligence, that helps computers to measure sentiment and determine which parts of language are significant.
Sarcasm detection is a specific case of sentiment analysis and presents a challenging task in natural language processing. Because sarcasm uses the opposite sentiment, it is hard to understand, not only for machines but also for humans. In this thesis, we test machine and deep learning methods for sarcasm classification in social media posts. Our two main approaches include BiLSTM models that take the response and various lengths of context into consideration. We conclude that too much context reduces the accuracy of the model. We observe that deep learning approaches yield better results than machine learning.

• Keywords: obrada prirodnog jezika; sarkazam; LSTM; kontekst; strojno učenje; duboko učenje; klasifikacija; društvene mreže
• Keywords (english): natural language processing; sarcasm; LSTM; context; machine learning; deep learning; classification; social media
• Language: english
• URN:NBN: urn:nbn:hr:168:289130
• Study programme: Title: 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 resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-01-14 21:47:54