search for: fmn2

Bad scaling will waste a lot of everyone's time. I put the data in a data frame mdat, then library(nlmrt) mdat<-read.csv("muzzamil.csv", header=T) fmn <- nlxb(y~a * (x^b), data=mdat, start=c(a=1,b=1), trace=T) fm <- nls(y~a * (x^b), data=mdat, start=c(a=1,b=1), trace=T) fmn2 <- nlxb(y~a2 * ((x-1979)^b2), data=mdat, start=c(a2=1,b2=1), trace=T) fm2 <- nls(y~a2 * ((x-1979)^b2), data=mdat, start=c(a2=1,b2=1), trace=T) found a possible solution quickly with the last 2 forms. Without adjusting the x data, you get a very ill-conditioned problem. JN > Message: 2...