• JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
 
  Bookmark and Share
 
 
Master's Dissertation
DOI
https://doi.org/10.11606/D.55.2022.tde-06092022-100818
Document
Author
Full name
Bruno Aparecido Barbosa
E-mail
Institute/School/College
Knowledge Area
Date of Defense
Published
São Carlos, 2022
Supervisor
Committee
Rezende, Solange Oliveira (President)
Carvalho, Veronica Oliveira de
Delbem, Alexandre Cláudio Botazzo
Lobato, Fábio Manoel França
Title in Portuguese
Predição do movimento de ações da Petrobras a partir de notícias
Keywords in Portuguese
Classificação
Mineração de Textos
Notícias
Petrobras
Abstract in Portuguese
Com a crescente entrada de novos investidores no mercado acionário brasileiro vê-se uma crescente circulação de conteúdos referentes a empresas, oportunidades de investimentos, avaliações e fatos relevantes sobre investimentos do mercado financeiro brasileiro. Esse movimento, somado a possibilidade de veiculação de conteúdos na internet torna difícil para os investidores assimilar novos conteúdos que são veiculados diariamente visando a melhor tomada de decisão na alocação de recursos para seus investimentos. Uma saída para abranger mais conteúdos referentes a alguma possibilidade de investimento e selecionar somente o conteúdo mais relevante é a utilização de algoritmos que possam extrair de um elevado conjunto de informações as mais relevantes para o interesse do investidor. O uso de tecnologias de aprendizado de máquina vem sendo usado em diversas áreas do conhecimento, dentre elas a classificação ou previsão do movimento do mercado financeiro utilizando notícias. No entanto, até agora poucos trabalhos utilizam modelos de linguagem natural baseados em redes neurais, como text embedding. Neste trabalho, é proposto a utilização do modelo BERT (Bidirectional Encoder Representations from Transformers) em comparação com o modelo BoW (Bag of Words) com o classificador Naive Bayes SVM, mais comumente utilizado para a finalidade de prever o movimento de uma ação do mercado financeiro através de notícias. Será utilizado notícias de jornais brasileiros separando somente as notícias referentes a empresa Petrobras
Title in English
Petrobras share movement prediction based on news
Keywords in English
Classification
News
Petrobras
Text Mining
Abstract in English
With the increasing entry of new investors in the Brazilian stock market, there is a growing circulation of content referring to companies, investment opportunities, evaluations and relevant facts about investments in the Brazilian financial market, this movement, added to the possibility of broadcasting content on the internet makes it difficult for investors to assimilate new content that is broadcast daily, aiming at better decision-making in the allocation of resources for their investments. A way out to cover more content referring to some investment possibility and select only the most relevant content is the use of algorithms that can extract from a high set of information the most relevant to the investors interest. The use of machine learning technologies has been used in several areas of knowledge, among them the classification or prediction of the movement of the financial market using news. However, so far few works use natural language models based on neural networks, such as text embedding. In this work, we propose the use of the BERT model (Bidirectional Encoder Representations from Transformers) in comparison with the BoW model (Bag of Words) with the most commonly used Naive Bayes SVM classifier. for the purpose of predicting the movement of a stock in the financial market through news. News from Brazilian newspapers will be used, separating only the news referring to the company Petrobras
 
WARNING - Viewing this document is conditioned on your acceptance of the following terms of use:
This document is only for private use for research and teaching activities. Reproduction for commercial use is forbidden. This rights cover the whole data about this document as well as its contents. Any uses or copies of this document in whole or in part must include the author's name.
Publishing Date
2022-09-06
 
WARNING: Learn what derived works are clicking here.
All rights of the thesis/dissertation are from the authors
CeTI-SC/STI
Digital Library of Theses and Dissertations of USP. Copyright © 2001-2024. All rights reserved.