Izgradnja reprezentacija tekstova korisnika za predviđanje ličnosti na temelju strojnog učenja

master's thesis

• Title (english): Embedding Users' Texts for Machine Learning-Based Personality Prediction
• Author: Mate Mijolović
• Mentor: Jan Šnajder (mentor)
• Committee member: Jan Šnajder (predsjednik povjerenstva)
• Committee member: Marko Čupić (član povjerenstva)
• Committee member: Leonardo Jelenković (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2020-07-10, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Predviđanje osobnosti iz teksta atraktivno je istraživačko područje sa širokim mogućnostima primjene. Pouzdan pokazatelj osobnosti može nadopuniti naše znanje o osobi i poslužiti kao vrijedna značajka za razne sustave. Napredak u strojnom učenju i obradi prirodnog jezika otvorio je nove mogućnosti i pobudio interes istraživača. Noviji radovi na temu predviđanja osobnosti usmjereni su na korištenje novih skupova podataka s oznakama osobnosti, često iz perspektive analize teksta. Ovaj rad... More pristupa zadatku predviđanja osobnosti koristeći isključivo tekstualne podatke. Izgradili smo više automatiziranih sustava predviđanja osobnosti temeljenih na strojnom učenju i testirali njihove performanse na zadacima predviđanja osobnosti koristeći modele Big Five i Myers-Briggs Type Indicator. Provedeni eksperimenti pokazuju da je moguće predvidjeti osobnost iz teksta, iako je zadatak izazovan i pokazuje mnogo prostora za napredak. Less

Abstract (english)

Personality prediction from text is an attractive research area with broad application possibilities. Having a reliable indicator of an individual’s personality can complement our knowledge about the observed person and serve as a valuable feature for various decision-making systems. The recent advances in machine learning and natural language processing opened up new possibilities and aroused the interest of researchers. Most recent papers are focused on exploiting the new datasets... More with personality labels, often from the perspective of text analysis. This thesis approaches the task of personality prediction by using only text data to derive the input features. We build multiple automated personality prediction systems based on machine learning and evaluate their performance on the tasks of predicting the Big Five and Myers-Briggs Type Indicator personality traits. Our experiments indicate it is possible to perform personality prediction just by considering the text data, although the task is challenging and shows much room for improvement. Less

• Keywords: strojno učenje; duboko učenje; obrada prirodnog jezika; predviđanje osobnosti
• Keywords (english): machine learning; deep learning; natural language processing; personality prediction
• Language: english
• URN:NBN: urn:nbn:hr:168:523348
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-01-13 13:08:41