Towards Unified Multimodal Financial Forecasting: Integrating Sentiment Embeddings and Market Indicators via Cross-Modal Attention
We propose {STONK} (Stock Optimization using News Knowledge), a multimodal framework integrating numerical market indicators with sentiment-enriched news embeddings to improve daily stock-movement prediction. By combining numerical \& textual embeddings via feature concatenation and cross-modal attention, our unified pipeline addresses limitations of isolated analyses. Backtesting shows {STONK} outperforms numeric-only baselines. A comprehensive evaluation of fusion strategies and model configurations offers evidence-based guidance for scalable multimodal financial forecasting. Source code is available on {GitHub}11https://github.com/sarthak-12/thesis-dsaa/.
- Published in:
IEEE 12th International Conference on Data Science and Advanced Analytics (DSAA) - Type:
Inproceedings - Authors:
- Year:
2025 - Source:
https://ieeexplore.ieee.org/document/11247993
Citation information
: Towards Unified Multimodal Financial Forecasting: Integrating Sentiment Embeddings and Market Indicators via Cross-Modal Attention, IEEE 12th International Conference on Data Science and Advanced Analytics (DSAA), 2025, 1--6, October, https://ieeexplore.ieee.org/document/11247993, Khanna.etal.2025b,
@Inproceedings{Khanna.etal.2025b,
author={Khanna, Sarthak; Berger, Armin; Berghaus, David; Deusser, Tobias; Sparrenberg, Lorenz; Sifa, Rafet},
title={Towards Unified Multimodal Financial Forecasting: Integrating Sentiment Embeddings and Market Indicators via Cross-Modal Attention},
booktitle={IEEE 12th International Conference on Data Science and Advanced Analytics (DSAA)},
pages={1--6},
month={October},
url={https://ieeexplore.ieee.org/document/11247993},
year={2025},
abstract={We propose {STONK} (Stock Optimization using News Knowledge), a multimodal framework integrating numerical market indicators with sentiment-enriched news embeddings to improve daily stock-movement prediction. By combining numerical \& textual embeddings via feature concatenation and cross-modal attention, our unified pipeline addresses limitations of isolated analyses. Backtesting shows...}}