文件列表:

Cvariances.c (1872, 2005-05-09)
farimasim.R (2353, 2005-06-27)
fdsim.f (113884, 2005-06-27)
fft-fgn.R (2197, 2005-07-14)
locwhitt.R (2564, 2005-06-27)
mCrs.c (3555, 2005-05-09)
NcalcHurst.R (3396, 2005-07-14)
plotper.R (7755, 2005-06-26)
plotrs.R (18847, 2005-06-26)
plotvar.R (8813, 2005-06-26)
wavedecomp.c (9832, 2005-06-26)
wavelet.R (20203, 2005-06-27)

These instructions were written for Ubuntu Linux but should work on any form of unix. The compiling line in particular will need adapting for windows systems (please refer to R documentation). The software here is almost entirely the work of other researchers and is credited appropriately in source code. Much of the software here can be found on the Comprehensive R network in the fseries package. First compile up the needed C code with: R CMD SHLIB Crs.c R CMD SHLIB Cvariances.c R CMD SHLIB wavedecomp.c R CMD SHLIB fdsim.f Start R > source("fft-fgn.R") > fgn<-fgngen(1000,0.75) Generates 1000 points of Fractional Gaussian Noise and stores in variable fgn > source("farimasim.R") > far<-farimagen(ar=c(0.2,0.1), ma=c (0.2,0.2), d= 0.2, n=1000) Generates 1000 points of FARIMA data with the AR and MA parameters as specified and H=0.7 (H= d+ 0.5) To calculate the Hurst parameter >source ("calcHurst.R") >calcHurst(far) >calcHurst(fgn) To remove a trend > x<-seq(1,length(far)) > pm<-lm(far ~ poly(x, degree=1)) > calcHurst(pm$residuals) To remove a poly order 10 (can take a while on long data setss > x<-seq(1,length(far)) > pm<-lm(far ~ poly(x, degree=10)) > calcHurst(pm$residuals)

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：