This paper presents a novel approach for forecasting stock prices. Specifically, the approach consists of an ensemble of various deep learning models, namely "multi-model". Each deep learning model produces its own forecast, then all the forecasts are combined into a unique one, according to different strategies and depending on different error metrics. The final forecasts provided by the multi-model have resulted in more reliable predictions than those provided by the individual deep learning models.

Pellattiero, D., Candelieri, A. (2024). Multi-model forecasting for finance. In M. Corazza, F. Gannon, F. Legros, C. Pizzi, V. Touzé (a cura di), Mathematical and Statistical Methods for Actuarial Sciences and Finance MAF2024 (pp. 248-254). Springer Nature [10.1007/978-3-031-64273-9_41].

Multi-model forecasting for finance

Candelieri A.
2024

Abstract

This paper presents a novel approach for forecasting stock prices. Specifically, the approach consists of an ensemble of various deep learning models, namely "multi-model". Each deep learning model produces its own forecast, then all the forecasts are combined into a unique one, according to different strategies and depending on different error metrics. The final forecasts provided by the multi-model have resulted in more reliable predictions than those provided by the individual deep learning models.
Capitolo o saggio
CEEMDAN; Deep learning; Forecasting; Neural networks; Recurrent neural networks; Stock markets; Wavelet thresholding;
English
Mathematical and Statistical Methods for Actuarial Sciences and Finance MAF2024
Corazza, M; Gannon, F; Legros, F; Pizzi, C; Touzé, V
2024
9783031642722
Springer Nature
248
254
Pellattiero, D., Candelieri, A. (2024). Multi-model forecasting for finance. In M. Corazza, F. Gannon, F. Legros, C. Pizzi, V. Touzé (a cura di), Mathematical and Statistical Methods for Actuarial Sciences and Finance MAF2024 (pp. 248-254). Springer Nature [10.1007/978-3-031-64273-9_41].
none
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/10281/551724
Citazioni
  • Scopus 0
  • ???jsp.display-item.citation.isi??? 0
Social impact