 Unofficial Call and Response lyrics
 hidden highlight selection when search with * or # is active
 freestanding quarter turn stair
 Who was the first to say “Shut up and calculate!”?
 Anatomically Correct Demons
 The gun which makes no sound
 Do Gods inherit sins committed by their avataras?
 Can the list of Divyadesams be added to or subtracted from?
 Were Periyalwar and Andal reborn as cowherds?
 Required alt text field in assets
 Issue with Matrix `type` property and Live Preview?
 Help me with this rebus puzzle please
 Who am I?  Riddles
 What and where can I be?
 Bongard number 2 is here
 Wherever I May Roam 2000
 Stuck in an Appliance
 Please help with this geocaching puzzle
 What are the main benefits and issues surrounding data today?
 Using Random Forest Probabilities for customer sensitivity
Accuracy for GARCH models
How does one calculate the accuracy of forecasts given by GARCH models considering GARCH is run on returns. Assuming GARCH is a derivative of a regression based prediction model, would regular statistics like R squared, MAPE/ SMAPE etc be the right indicator for the performance? Unlike ARIMA where the predictive power just dies down after a forecast interval, I experience GARCH forecasting values of almost any time period specified. How would one be able to identify if there is any randomness in the forecasted values?
The best way to check the accuracy of a Garch model is to use the methodology of Hansen and Lunde (2005). In this paper they actually compared the accuracy of 330 Archtype models and concluded that Garch(1,1) was superior in their sample.
The paper describes at great length the way to do it. But in a nutshell:
Estimate archtype your model in monthly data using a window from $[t_{start}, t_{end}$. Compute volatility forecast for month $t_{end+1}

The best way to check the accuracy of a Garch model is to use the methodology of Hansen and Lunde (2005). In this paper they actually compared the accuracy of 330 Archtype models and concluded that Garch(1,1) was superior in their sample.
The paper describes at great length the way to do it. But in a nutshell:
Estimate archtype your model in monthly data using a window from $[t_{start}, t_{end}$. Compute volatility forecast for month $t_{end+1}$.
Compute realized volatility during month $t_{end+1}$ using daily data (e.g. sum of squared returns).
Subtract your forecasted volatility from realized volatility and square it. Do this for several months and check the sum of squared residuals.
You can do this exercise using many frequencies. My example above is for monthly frequency. Just replace the appropriate definitions.
20180315 17:42:47