Guardado en:
| Autores principales: | , , , |
|---|---|
| Formato: | Recurso digital |
| Lenguaje: | |
| Publicado: |
Zenodo
2025
|
| Acceso en línea: | https://doi.org/10.5281/zenodo.15401754 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
Tabla de Contenidos:
- <p> Abstract—Stock price prediction is a complex challenge in<br>f<br> luenced by various factors, including historical trends and<br> market sentiment. Existing models often rely on single-source<br> data, limiting their forecasting accuracy. This study presents<br> a hybrid approach that integrates Long Short-Term Memory<br> (LSTM) networks for time-series forecasting with sentiment<br> analysis derived from financial news. To dynamically combine<br> the two prediction outputs, Bayesian Optimization is employed<br> to calculate optimal weights, ensuring adaptability to market<br> f<br> luctuations. Experimental results demonstrate that the proposed<br> method achieves improved accuracy, reducing the Mean Squared<br> Error compared to traditional fixed-weight models. This research<br> contributes to the field of financial forecasting by providing a<br> more responsive and data-driven prediction framework</p>