Description

We modeled the logarithm of the daily increase in closing prices of the stocks of Astrazeneca, Johnson&Johnson, Moderna and Pfizer on the yearly period ending on April 25th, 2021.

In particular we used:

• Augmented Dickey Fuller test
• ARMA modelling
• Model selection with AIC, BIC
• One-sample Kolmogorov-Smirnov non-parametric test for normality
• Ljung-Box Q-test
• GARCH modelling
• EGARCH modelling
• Comparison of forecasting performances with MAE and MSE

Project report

The full report of the project, containing all the references and a complete discussion of the modelling choices is available in Project Description.pdf.

Authors

Mirko Giovio Politecnico di Milano - MSc in Statistical Learning

Daniele Venturini Politecnico di Milano - MSc in Statistical Learning

Paolo Vergottini Politecnico di Milano - MSc in Statistical Learning

Acknowledgments

The projects were carried out during Econometrics course of PoliMi, 2021. Special thanks to Prof. Sgarra, under whose guidance all the material was developed.