I am working with the splm package. I use the spgm function: general estimation of a panel data model. Based on this approach, I know it is possible to compute a R2, eg the ratio of variation explained by a given model.
My model is : bivmod<-spgm(logIKA~NBLITRE0+NBLITRE1,data=mydatap,listw=comsKnn.nbW,spatial.error=TRUE) I know that we can calculate the R^2 as the variance of the fitted values from the reduced form of the model (Yfitted) over the variance of y (here logIKA). Since I am using the option lag = FALSE; the fitted value for an error model are Yfitted = X\beta. If I well understood : *1) We can get X* (NT * k matrix of observations on the non-stochastic regressors. with N: spatial units , T: time unit et k : number of non-stochastic regressors) with ** >bivmod$model[,c(2,3)] # in the matrix of the data used (bivmod$model), the first column corresponds to the response variable and the subsequent one to regressors). NBLITRE0 NBLITRE1 ADAM-LES-PASSAVANT-2004 0 0 ADAM-LES-PASSAVANT-2005 0 0 ADAM-LES-PASSAVANT-2006 0 0 ADAM-LES-PASSAVANT-2007 0 0 ADAM-LES-PASSAVANT-2008 0 0 ADAM-LES-VERCEL-2004 0 0 *2) Beta* corresponds to coefficients associated to each regressor : we can get them with ** >bivmod$coefficients NBLITRE0 NBLITRE1 -0.001131662 -0.001083650 But I do not know what means the " \ " in the formula. I suppose it is different from "/" which means "divide by ". Any hint appreciated. Regards, Marion -- Marion Jacquot Laboratoire de Chrono-environnement UMR UFC/CNRS 6249 USC INRA Université de Franche-Comté Place Leclerc F-25030 Besançon cedex FRANCE Tel. : +33 (0)381 665 829 Fax : +33 (0)381 665 797 http://chrono-environnement.univ-fcomte.fr/ [[alternative HTML version deleted]]
______________________________________________ R-help@r-project.org mailing list 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.
