Home // SOTICS 2023, The Thirteenth International Conference on Social Media Technologies, Communication, and Informatics // View article

Stock Price Prediction Based on Investor Sentiment Using BERT and Transformer Models

Authors:
Chien-Cheng Lee
Anish Sah

Keywords: Bert; Transformer; StockTwits; Stock Returns

Abstract:
This paper investigates the impact of investor sentiment on the stock market by predicting stock closing prices and future trends in stock returns. Our study involves gathering abundant investor messages from three social media platforms: Stocktwits, Yahoo Finance, and Reddit. To gauge investor sentiment from the collected messages, we employ Bidirectional Encoder Representations from Transformers (BERT), a transformer-based pre-trained language model. We present a novel application of a Transformer-based model for stock trend prediction. This model architecture leverages the self-attention mechanism to capture the interdependence of stock data, facilitating accurate forecasting of stock trends. By integrating investor sentiment with stock prices and inputting this combined information into the transformer model, we predict the performance of APPLE and SPY stocks datasets. Our experimental results reveal that the transformer model exhibits strong performance regardless of whether sentiment features are included. Moreover, incorporating sentiment does enhance the forecasting accuracy for both stock closing prices and future trends in stock returns.

Pages: 1 to 5

Copyright: Copyright (c) IARIA, 2023

Publication date: November 13, 2023

Published in: conference

ISSN: 2326-9294

ISBN: 978-1-68558-103-9

Location: Valencia, Spain

Dates: from November 13, 2023 to November 17, 2023