Understand and Model Cryptocurrencies Volatility Using GARCH Variants

oxboz

New Member
Hi everyone, I have been working on modeling cryptocurrencies volatility using GARCH and its variants. I have run autocorrelations on squared daily log return and realized variance (intraday data both 1-min and 5-min). The lag is quite large (40+). When I used the default values provided with ARCH package (p=1, q=1), the models seem to fit okay based on MAE, MAPE and RMSE.

My apologies for this long post.
Understand and Model Cryptocurrencies Volatility Using GARCH Variants

I have included all calculations and charts while raising the question in the middle of this post. Please let me know if you have suggestions/comments. Much appreciated.


squared_daily_log_return_acf.png

5_min_daily_realized_variance_acf.png
1_min_daily_realized_variance_acf.png
 

Attachments

Top