r - Panel data regression for large datasets (ffdf) -
i using ffdf class store panel dataset quite big in size. know if there way of running fixed effects regression data in class.
library(ffbase) size <- 100000 index <- 100 <- ffrandom(n=index, rfun = runif) panel_index <- ffseq_len(index) times_vec <- ffrep.int(size,times=index) b <- ffrep.int(a, times=times_vec) panel_index <- ffrep.int(panel_index, times=times_vec) error <- ffrandom(n=size*index, rfun=rnorm) intercept <- ffrep.int(10, times=size*index) x <- ffrandom(n=size*index, rfun=rnorm) y <- x+intercept+b+error dataf <- ffdf(y = y, x = x, panel_index = panel_index) ## on disk dim(dataf)
i run like
library(plm) fixed <- plm(y ~ x, data=dataf, index="panel_index", model="within")
but following error
error in is.finite(x[[i]]) : default method not implemented type 'list'
what alternative packages can handle ffdf data?
Comments
Post a Comment