Predikcija osobnosti na temelju obrade prirodnog jezika uporabom metoda dubokog učenja

master's thesis

• Title (english): Personality prediction based on natural language processing using deep learning methods
• Author: Marija Frković
• Mentor: Boris Vrdoljak (mentor)
• Committee member: Boris Vrdoljak (predsjednik povjerenstva)
• Committee member: Mario Brčić (član povjerenstva)
• Committee member: Mirta Baranović (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2020-07-07, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

U sklopu ovog diplomskog rada istraženi su osnovni koncepti obrade prirodnog jezika te je nad skupom podataka, koji sadrži tekstualne objave i Myers-Briggs tipove osobnosti autora, izvršen niz radnji za obradu teksta. Za zadatke obrade prirodnog jezika najčešće se koriste LSTM neuronske mreže, dizajnirane za obradu duljih nizova podataka. U sklopu rada je istražen utjecaj strukturalnih hiperparametara na rezultate klasifikacije teksta pomoću LSTM neuronskih mreža. Zatim je istražen utjecaj primjene troškovno osjetljivog učenja kao metode rješavanja problema neuravnoteženosti skupa podataka. Dobiveni rezultati su iskorišteni za izradu aplikacije koja kao ulaz prima tekstualnu datoteku, a vraća MBTI tip osobnosti autora teksta dobiven pomoću odabranih modela neuronske mreže.

Abstract (english)

As part of this thesis, the basic concepts of natural language processing have been explored and a series of word processing operations have been performed on the dataset, which consists of textual posts and Myers-Briggs personality types of the authors. LSTM neural networks, designed for processing long sequences of data, are most commonly used for natural language processing tasks. In this thesis, the influence of structural hyperparameters on the results of text classification using LSTM neural network has been analyzed. Then, the influence of the application of cost-sensitive learning as a method of solving the problem of dataset imbalance has also been analyzed. The obtained results were used to create an application that accepts text file as an input and outputs the MBTI personality type of the author of the text using selected neural network models.

• Keywords: obrada prirodnog jezika; Myers-Briggs tipovi osobnosti; LSTM neuronske mreže; klasifikacija teksta; strukturalni hiperparametri; troškovno osjetljivo učenje
• Keywords (english): natural language processing; Myers-Briggs personality types; LSTM neural networks; text classification; structural hyperparameters; cost-sensitive learning
• Language: croatian
• URN:NBN: urn:nbn:hr:168:025541
• 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 12:15:48