Title: Financial sentiment analysis of news articles with long text corpus for equity portfolio construction

Authors: Senthil Arasu Balasubramanian; J. Nancy Christina; P. Sridevi

Addresses: Department of Management Studies, National Institute of Technology, Tiruchirappalli – 620015, Tamilnadu, India ' Department of Management Studies, National Institute of Technology, Tiruchirappalli – 620015, Tamilnadu, India ' Department of Management Studies, National Institute of Technology, Tiruchirappalli – 620015, Tamilnadu, India

Abstract: Forecasting stock performance is a well-researched area. In recent times, textual data related to stock market are considered to have more meaningful insights and various natural language processing (NLP) techniques are employed to process it. Several studies have used news headlines to predict stock market performance and most of the studies focus on short-term forecasting which considers lags of days or less and pose a higher risk than investing over a long-term. In this study, long text corpus of news articles of non-financial stocks from Nifty 50 is analysed with an objective to construct an equity portfolio. A pre-trained NLP model FinBERT was used to analyse the sentiment of the financial text. The resulting portfolio was found to outperform the market. However, few stocks with extraordinary performance were missed as the stocks for the portfolio were selected using news articles as the only source.

Keywords: sentiment analysis; natural language processing; NLP; news corpus; BERT-base-NER; FinBERT; Nifty; equity portfolio.

DOI: 10.1504/IJICBM.2025.144909

International Journal of Indian Culture and Business Management, 2025 Vol.34 No.3, pp.389 - 407

Received: 28 Jul 2023
Accepted: 01 Aug 2023
Published online: 07 Mar 2025 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article