Título: Applying LSTM for Stock Price Prediction with Sentiment Analysis
Autores: Alexandre Heiden and Rafael Parpinelli.
Resumo:
Financial news has been proven to be valuable source of information for the evaluation of stock market volatility. Most of the attention has been given to social media platforms, while news from vehicles such as newspapers are not as widely explored. Newspapers provide, although in a smaller volume, more reliable information than social media platforms. In this context, this research aims to examine the influence of financial news within the stock price prediction problem, by using the VADER sentiment analysis model to process the news and feed the sentiments as a feature into a LSTM-based stock price prediction model, along with the historical data of the assets. Experiments indicate that the model has better results when the news’ sentiments are considered, and the model demonstrates potential to accurately predict stock prices up to around 60 days into the future.
Palavras-chave:
Stock price prediction, Sentiment analysis, Financial news.
Páginas: 8
Código DOI: 10.21528/CBIC2021-45
Artigo em pdf: CBIC_2021_paper_45.pdf
Arquivo BibTeX: CBIC_2021_45.bib