Generalized AutoRegressive Conditional Heteroskedasticity ( Volatility Model )

Conclusion

We created a Python class garchOneOne that allows fitting a GARCH(1,1) process to financial series. Our estimations are coherent, for both the S&P 500 and CAC 40 indices, to the arch_model fitting from the arch package. Our model gives a good approximation of the behavior of uncertainty in the financial markets. However, we tend to underestimate the long-term variance parameter of the GARCH process.