wgtmethod in correlation

For questions regarding programming in the EViews programming language.

Moderators: EViews Gareth, EViews Jason, EViews Moderator, EViews Matt

styx
Posts: 11
Joined: Mon Oct 21, 2013 5:59 am

wgtmethod in correlation

Postby styx » Thu Jan 09, 2014 9:33 am

Hello,

I'd like to account for possible heteroskedasticity in my series when I calculate correlation coefficients. There is a 'wgtmethod' option in the 'cor' 'interactive use command'. My question would be what is this option? Is it a kind of GLS option? How can be activated this option?

When I'd like to apply it to a group, it does not work.
For instance let grp be a group for which I'd like to calculate a correlation matrix. If I try grp.cor(wgtmethod="stdev"), it calculates the ordinary correlation matrix. I've also tried without "", the result is the same.

Could you help me?
Thank you.
Best,
Peter

EViews Glenn
EViews Developer
Posts: 2672
Joined: Wed Oct 15, 2008 9:17 am

Re: wgtmethod in correlation

Postby EViews Glenn » Thu Jan 09, 2014 2:37 pm

The weights are row-wise (observation weights). You need to supply a weight series. The weight-method corresponds to the coding of the weight series. The option that you specified indicates that the weights are the standard-deviations of the observation. The correlations will then be computed using the appropriate transformation to heterogeneity (in this case dividing the series by the weights).


Return to “Programming”

Who is online

Users browsing this forum: No registered users and 23 guests