Gundermann, M Wandrille (2024) Kernels Methods applied to financial time series forecasting PRE - Projet de recherche, ENSTA.
Fichier(s) associé(s) à ce document :
| PDF 1981Kb |
Résumé
Studies on the behaviour of financial markets very often utilize regressions in the analysis process. However, the vast majority of these are linear regressions. For example, a firm’s excess returns are calculated by taking the difference between the actual return and the expected return, which has itself been calculated using linear regression via the CAPM method. All these econometric variables are modelled using a linear combination of time series, but there is no theoretical guarantee of a linear relationship between them and their parameters. It is therefore relevant to wonder whether non-linear models – which are therefore more complex to implement – are able to approximate these variables better in order to increase the accuracy of the findings of these studies on the behaviour of financial markets.
Type de document: | Rapport ou mémoire (PRE - Projet de recherche) |
---|---|
Mots-clés libres: | Forecast Returns |
Sujets: | Sciences de l'économie, de la gestion et de la société Mathématiques et leurs applications |
Code ID : | 10103 |
Déposé par : | Wandrille GUNDERMANN |
Déposé le : | 03 sept. 2024 09:56 |
Dernière modification: | 03 sept. 2024 09:56 |