Roger Koenker
2015-Jun-11 13:33 UTC
[R] Quantile regression model with nonparametric effect and interaction
The main effect trend seems rather dangerous, why not just estimate the f?s in a loop? url: www.econ.uiuc.edu/~roger Roger Koenker email rkoenker at uiuc.edu Department of Economics vox: 217-333-4558 University of Illinois fax: 217-244-6678 Urbana, IL 61801> On Jun 11, 2015, at 1:57 AM, Waltl, Sofie (sofie.waltl at uni-graz.at) <sofie.waltl at uni-graz.at> wrote: > > Dear all, > > I would like to estimate a quantile regression model including a bivariate nonparametric term which should be interacted with a dummy variable, i.e., > log p ~ year + f(a,b):year. > I tried to use Roger Koenker's quantreg package and the functions rqss and qss but it turns out that interactions are not possible in this setting. Also weights are not implemented yet to build a work-around. I am looking for something like the by-statement in Simon Wood's package mgcv. Does anything comparable exist? > I am grateful for any help on this issue. > > Kind regards, > S. Waltl > > > [[alternative HTML version deleted]] > > ______________________________________________ > R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code.
Waltl, Sofie (sofie.waltl@uni-graz.at)
2015-Jun-11 14:05 UTC
[R] Quantile regression model with nonparametric effect and interaction
The idea is to move from regional dummies interacted with time dummies (model 1) to a smooth spline (defined on longitudes and latitudes) interacted with time dummies (model 2), i.e., Model 1: Log p ~ X\beta + REGION*YEAR Model 2: Log p ~ X\beta + f(long, lat)*YEAR Estimating the f's in a loop therefore does not really help... -----Original Message----- From: Roger Koenker [mailto:rkoenker at illinois.edu] Sent: Donnerstag, 11. Juni 2015 15:33 To: Waltl, Sofie (sofie.waltl at uni-graz.at) Cc: r-help at r-project.org Subject: Re: [R] Quantile regression model with nonparametric effect and interaction The main effect trend seems rather dangerous, why not just estimate the f's in a loop? url: www.econ.uiuc.edu/~roger Roger Koenker email rkoenker at uiuc.edu Department of Economics vox: 217-333-4558 University of Illinois fax: 217-244-6678 Urbana, IL 61801> On Jun 11, 2015, at 1:57 AM, Waltl, Sofie (sofie.waltl at uni-graz.at) <sofie.waltl at uni-graz.at> wrote: > > Dear all, > > I would like to estimate a quantile regression model including a > bivariate nonparametric term which should be interacted with a dummy variable, i.e., log p ~ year + f(a,b):year. > I tried to use Roger Koenker's quantreg package and the functions rqss and qss but it turns out that interactions are not possible in this setting. Also weights are not implemented yet to build a work-around. I am looking for something like the by-statement in Simon Wood's package mgcv. Does anything comparable exist? > I am grateful for any help on this issue. > > Kind regards, > S. Waltl > > > [[alternative HTML version deleted]] > > ______________________________________________ > R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide > http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code.