| lazy {RTisean} | R Documentation |
Tools for performing nonlinear noise reduction.
nrlazy(series, l, x = 0, c = 1, m = 5, d = 1, i = 1, r, v) lazy(series, m, r, v, i = 1, l, x = 0, c = 1)
series |
a vector or a matrix. |
m |
embedding dimension. |
d |
delay for the embedding. |
r |
neighborhood size. |
v |
neighborhood size as fraction of data standard deviation. |
i |
number of iterations. |
l |
number of values to be read. |
x |
number of values to be skipped. |
c |
column to be read. |
In nrlazy each embedded point is replaced by the average vector
calculated in its neighborhood with a given size.
In lazy only the central component of each vector is corrected.
A vector containing the filtered time series.
T. Schreiber, Extremely simple nonlinear noise reduction method, Phys. Rev. E 47, 2401 (1993)
## Not run:
par(mfrow=c(1,2))
n <- 5000
dat <- henon(n)
dat[,1] <- dat[,1] + rnorm(n)/25
delayeddat <- embed(dat,d=2)
filtereddat <- lazy(dat,m=5,v=0.05,c=2)
delayedfiltdat <- embed(filtereddat,d=2)
plot(delayeddat, cex= 0.2, cex.main=0.7, xlab="", ylab="", main = "Embedded noisy Henon attractor filtered with lazy")
points(delayedfiltdat, cex=0.2, col=2)
legend(-1,-0.5, c("Noisy data","Filtered data"),fill=c(1,2),bty="n",cex=0.7)
filteredat2 <- nrlazy(dat,m=5,v=0.2)[,1]
delayedfiltdat2 <- embed(filteredat2,d=2)
plot(delayeddat, cex= 0.2, cex.main=0.7,xlab="", ylab="", main = "Embedded noisy Henon attractor filtered with nrlazy")
points(delayedfiltdat2 , cex=0.2, col=2)
legend(-1,-0.5, c("Noisy data","Filtered data"),fill=c(1,2),bty="n",cex=0.7)
## End(Not run)