| Titre : |
Learning by filtering : a stock price forecasting competition using state-space models |
| Type de document : |
Mémoire |
| Auteurs : |
Camilo VELASCO FRANCO, Auteur |
| Année de publication : |
2021 |
| Importance : |
57 p. |
| Note générale : |
Pour accéder aux fichiers PDF, merci de vous identifier sur le catalogue avec votre compte Office 365 via le bouton CONNEXION en haut de page. |
| Langues : |
Anglais (eng) |
| Mots-clés : |
Management
ACTION ; COMPETITIVITE ; PREVISION ; PRIX
|
| Résumé : |
Forecasting competitions draw the attention of both researchers and professionals. In fact, they have proven their pertinence by challenging different methods with multiple time series from a wide range of sectors and different frequencies, unveiling the capability of such methods to anticipate the movements of the studied time series. There have been controversies on whether these applications provide beneficial results or not, but what is true is that these competitions have played an essential role in the developments made in the predictive data science field. In this research, a forecasting competition is proposed exclusively using around 1 697 time series from stocks of 11 world indices since 2007. The competitors are known benchmarks such as the naïve method, but also proprietary methods developed by Prof. Giacomo Sbrana and other widely known ones, most of which were presented using a state space representation in the MSOE world applying the Kalman Filters. The results of this competition show, in a general and detailed way, the performance of the different methods and specific information on which method is best fitted for predicting specific asset prices. Moreover, the competition provides the accuracy of predicting single prices and the accuracy of predicting the direction. This information could be handy in the investment process. In conclusion, there is proof that competitions provide exciting and valuable results suggesting that assets behave differently. Thus their price movements can be anticipated by applying methods that are not necessarily the conventional ones |
| Programme : |
MS Analyse Financière Internationale (ft)- Reims |
| Permalink : |
https://cataloguelibrary.neoma-bs.fr/index.php?lvl=notice_display&id=539126 |
|  |
Learning by filtering : a stock price forecasting competition using state-space models / 2021 
