FinBERT-Based Sentiment Integration in Hybrid CNN– BiLSTM Models For Stock Price Forecasting

Mohammad Tyas Pawitra; Lukman Abdurrahman; Hanif Fakhrurroja

doi:10.15408/jti.v19i1.49466

Authors

Mohammad Tyas Pawitra Information System Study Program, School of Industrial Engineering, Telkom University
Lukman Abdurrahman Information System Study Program, School of Industrial Engineering, Telkom University
Hanif Fakhrurroja Information System Study Program, School of Industrial Engineering, Telkom University

DOI:

https://doi.org/10.15408/jti.v19i1.49466

Keywords:

stock prediction, deep learning, sentiment analysis, CNN, LSTM, BiLSTM, CNN-BiLSTM

Abstract

This study investigates sentiment-aware deep learning models for short-term stock price forecasting using NVIDIA (NVDA) as a representative high-volatility technology stock. Four architectures—CNN, LSTM, BiLSTM, and a hybrid CNN–BiLSTM—are evaluated under two configurations: without sentiment and with FinBERT-based financial news sentiment integrated as a continuous contextual feature. Historical OHLV data are combined with sentiment information to enable multimodal learning under a controlled experimental setting. The results demonstrate that recurrent architectures consistently outperform convolution-only models, highlighting the importance of temporal dependency modeling in financial time series. Among all configurations, the hybrid CNN–BiLSTM with FinBERT sentiment achieves the best overall performance, yielding the highest R², the lowest MAE and RMSE, and the smallest overfitting gap. Bootstrap-based confidence intervals indicate stable generalization, while Wilcoxon signed-rank tests confirm that the observed performance improvements are statistically significant. The study also presents a near real-time deployment framework with low inference latency, demonstrating practical applicability for decision-support systems. Overall, the findings show that effective alignment between local feature extraction, bidirectional temporal modeling, and contextual sentiment integration is critical for improving stock price. forecasting accuracy and robustness.

References

[1] W. Zhang and S. Liu, “The impact of news on financial markets: Evidence from sentiment analysis,” Finance Research Letters, 2022.

[2] T. Chen and H. Zhao, “Stock Market Prediction Based on Text Sentiment Analysis,” IEEE Access, 2021.

[3] J. Kharpal, “Nvidia becomes world’s most valuable chipmaker amid AI boom,” CNBC, 2023.

[4] M. Arora and J. Singh, “Volatility determinants in technology stocks,” Journal of Economic Studies, 2022.

[5] G. Box and G. Jenkins, Time Series Analysis: Forecasting and Control, Wiley, 2016.

[6] Y. Fang et al., “A survey of deep learning in financial market prediction,” IEEE Transactions on Neural Networks and Learning Systems, 2021.

[7] S. Hochreiter and J. Schmidhuber, “Long Short-Term Memory,” Neural Computation, 1997.

[8] C. Lee and J. Yoo, “CNN–LSTM Hybrid Networks for Stock Price Prediction,” Expert Systems with Applications, 2020

[9] S. Aluvala et al., “Hybrid CNN-BiLSTM model for time series forecasting,” IEEE Access, 2023.

[10] X. Li et al., “News-Driven Stock Prediction Using Deep Learning,” Information Sciences, 2020..

[11] N. Ranco et al., “The Effects of Twitter Sentiment on Stock Price Behavior,” Journal of Big Data, 2015.

[12] C. Hutto and E. Gilbert, “VADER: A rule-based model for sentiment analysis,” ICWSM, 2014.

[13] A. N. Ma’aly, D. Pramesti, A. D. Fathurahman, and H. Fakhrurroja, “Exploring sentiment analysis for the Indonesian presidential election through online reviews using multi-label classification with a deep learning algorithm,” Information, vol. 15, no. 11, p. 705, 2024, doi: 10.3390/info15110705.

[14] D. Pramesti, H. Fakhrurroja, and R. K. M, “Public sentiment and GoTo stock price movement in Indonesia: A null-relationship study using Naïve Bayes and non-parametric measures,” Jurnal Teknik Informatika, vol. 18, no. 2, pp. 257–269, 2025, doi: 10.15408/jti.v18i2.46447.

[15] Y. Yang, L. Yang, and R. Qin, “FinBERT: A Pretrained Language Model for Financial Tasks,” arXiv:2006.08097, 2020.

[16] R. Huang et al., “Transformer-based sentiment models for financial forecasting,” Applied Soft Computing, 2022.

[17] M. Mohan et al., “Sentiment-aware deep learning models for stock prediction,” Procedia Computer Science, 2019.

[18] Y. Li and Y. Pan, “A survey of deep learning in financial market applications,” ACM Computing Surveys, 2021.

[19] Z. Zhao et al., “A Multimodal Deep Learning Framework for Market Movement Prediction,” IEEE Trans. on Knowledge and Data Engineering, 2022.

[20] P. Chapman et al., CRISP–DM 1.0: Step-by-Step Data Mining Guide, SPSS, 2000.

[21] I. Goodfellow, Y. Bengio, and A. Courville, Deep Learning, MIT Press, 2016.

[22] J. Shah, D. Vaidya, and M. Shah, “A comprehensive review on multiple hybrid deep learning approaches for stock prediction,” Intelligent Systems with Applications, vol. 16, p. 200111, 2022.

[23] M. T. Pawitra, H. Fakhrurroja and L. Abdurrahman, "Predicting Stock Market using CNN and BiLSTM Model," International Conference on Computer, Control, Informatics and its Applications (IC3INA), 2024.