Hello, I compute a singular spectrum ananlysis of a time series using ssa of the Rssa package. Then I compute the forecast based on the results of the singular spectrum ananlysis (ssa). Here I observe that the original time series and the forecast are discontinous. How can I force the forecast to start at the last value (x,y) of the original time series? This minimal setup should show the (my) problem library(Rssa) md=data.frame(time=1:2000,val=runif(1000)) sdd = ts(md[,2], start=0, freq=1) s<-ssa(sdd) f1 <- forecast(s,groups=list(1:4),len=60) plot(f1,xlim=c(1950,2100)) I use the latest version of Rssa, R on linux Many greets and TIA ingo