Time Series

what’s GARCH? 3_letters Wrote: ------------------------------------------------------- > Wow. Sorry for throwing ARMA out there guys. > > In other news…how much do you know about > GARCH?..hahahaha

MA/ARMA not part of LOSs. dont sweat it.

I noticed on the analyst forum website under the “Samples” section at the top of the screen there is a link to the Level 3 Quant section from Schwesser which covers time series. If anyone is interested, they have a very good very brief section on Moving Average models, ARMA, etc. These items are in the Level 2 curriculum although not covered in Schwesser. For anyone interested, while this isn’t specifically in an LOS for Level 2 it was not optional in the CFA curriculum. Basically, you can test to see if a model is MA by looking at the autocorrelations. MA models of order q will have the first q autocorrelations significant after which none are. AR models autocorrelations will gradually decrease. ARMA just combines an AR and a MA model but is highly unstable and bad for forecasting; you need at least 80 observations.

thanks joeisenb Wrote: ------------------------------------------------------- > I noticed on the analyst forum website under the > “Samples” section at the top of the screen there > is a link to the Level 3 Quant section from > Schwesser which covers time series. If anyone is > interested, they have a very good very brief > section on Moving Average models, ARMA, etc. > These items are in the Level 2 curriculum although > not covered in Schwesser. For anyone interested, > while this isn’t specifically in an LOS for Level > 2 it was not optional in the CFA curriculum. > Basically, you can test to see if a model is MA by > looking at the autocorrelations. MA models of > order q will have the first q autocorrelations > significant after which none are. AR models > autocorrelations will gradually decrease. ARMA > just combines an AR and a MA model but is highly > unstable and bad for forecasting; you need at > least 80 observations.