CODEWITHSUNDEEP
rgeospatial
user3910073
user3910073

Reputation: 531

Fitting variogram R

I have created a data frame like this

 head(df)
DISTBin   MSDBin
1       0  4.301515
2      20  5.599879
3      40  5.883236
4      60  6.430584
5      80  7.396374
6     100  7.561652

where the first column is a vector of distances (binned) and the second one is a vector of mean square differences for a given quantity.

I would like then to fit df$MSDBin ~ df$DISTBin using a Gaussian model.

I cannot use gstat or similar since I don't have a variogram object (I had to compute the variogram by myself to discard pairs from the same ID).

Any help?

Upvotes: 1

Views: 154

Answers (1)

user3910073
user3910073

Reputation: 531

I have found the answer myself.

Gau <- function(h,tau,sigma.sq,range){
    return(tau + sigma.sq * (1-exp(-(h*h)/(range*range))))
}

Gau_Fit <- nls(MSDBin ~ Gau(DISTBin,tau,sigma.sq,range),
                 data = df,
                 start = list(tau = 0, sigma.sq = 10, range = 1000))

Upvotes: 2

Related Questions