Abstract
The COVID-19 pandemic offers a unique opportunity to examine the role of Internet signals in financial markets, specifically during the peak period when rapid dissemination of news about emerging threats can overwhelm investors, making it crucial for them to find original signals from the Internet for their decision making. We study the effectiveness of integrating Internet signals with traditional datasets to predict stock market movements. We selected the pharmaceutical industry and the pandemic’s peak period to obtain pronounced and relevant signals from news portals and search engines, along with technical indicators for intra-day trading. We employed state-of-the-art models for tabular data, such as Random Forest, Gradient Boosting, and XGBoost, and plan to experiment with deep learning models next. Our findings indicate that incorporating Internet signals contributes to significantly better predictions of stock market movements. This study offers significant practical implications and contributes to machine learning-based research for business decision-making.
Recommended Citation
Dixit, Gaurav and Morankar, Satyam Kamalakar, "Decoding Internet Signals for Stock Market Movement: A Machine Learning Study on Pharma Sector During Covid-19" (2024). Proceedings of the 2024 Pre-ICIS SIGDSA Symposium. 24.
https://aisel.aisnet.org/sigdsa2024/24