Volatility Prediction in the Indian Share Market Using Sentiment Analysis

Title

Volatility Prediction in the Indian Share Market Using Sentiment Analysis

Creator

Gupta, Vansh; George, Jossy; Chanti, S.

Description

In addressing the challenge of accurate volatility prediction in the Indian share market, the study explores the performance of deep learning-based models using sentimentdriven features. A demo model was deployed using data from five major Nifty 50 stocks-RELIANCE, HDFCBANK, INFOSYS, ITC, and MARUTI-for the financial years 2020 to 2023. We compared the results of traditional ARIMA model and standalone LSTM and its hybrid variants: LSTM + CNN and LSTM+RNN. Sentiment scores were gathered from financial news using FinBERT and NLTK, and combined with stock price data to generate time-series features. While all models demonstrated promising results, the LSTM+RNN hybrid model consistently achieved the lowest MAE and RMSE, indicating improved learning of temporal dependencies. The standalone LSTM and LSTM+RNN models also showed positive results for Sharpe ratio and Maximum drawdown indicating strong economic significance of the models. The study emphasizes the potential of hybrid LSTM architectures in modeling market volatility driven by investor sentiment. Limitations included limited dataset size, exclusion of other volatility factors, and overfitting in early hybrid GARCH trials. Future work aims to expand data coverage, integrate hybrid GARCH models more effectively, and explore additional market indicators. This research highlights a scalable and effective approach for sentiment-informed volatility forecasting in financial domains. 2025 IEEE.

Source

Proceedings - 2025 International Conference on Transformative Computing Technologies, ICTCT 2025;pp.382-390

Date

01-01-2025

Publisher

Institute of Electrical and Electronics Engineers Inc.

Subject

ARIMA; Deep learning; FinBERT; Hybrid LSTM + RNN and LSTM + CNN; Indian share market; LSTM; MAE; RMSE; Sentiment analysis; Stock market; Volatility prediction

Coverage

Gupta V., CHRIST (Deemed to be University), Bengaluru, India; George J., CHRIST (Deemed to be University), Bengaluru, India; Chanti S., CHRIST (Deemed to be University), Bengaluru, India

Rights

Restricted Access; Hardcopy may be available in the library

Relation

ISBN: 979-833159195-3;

Format

online

Language

English

Type

Conference paper

Identifier

https://doi.org/10.1109/ICTCT69201.2025.00076

https://www.scopus.com/pages/publications/105037459775?origin=resultslist