Hi, I am trying the following: library(ismev) library(evd) fit <- gev.fit(x,show=FALSE) ks.test(x,pgev,fit$mle[1],fit$mle[2],fit$mle[3]) but I am getting: Warning message: cannot compute correct p-values with ties in: ks.test(x, pgev, fit$mle[1], fit$mle[2], fit$mle[3]) where x is: [1] 239 38 1 43 22 1 5 9 15 6 1 9 156 25 3 100 6 [18] 5 100 10 103 25 5 1 2 11 8 68 13 154 67 12 5 15 [35] 5 130 47 143 176 573 592 213 54 10 17 9 198 293 77 11 44 [52] 6 22 2 10 8 1 2 16 4 70 12 4 7 134 41 515 8 [69] 200 169 2 13 49 218 48 34 74 19 44 128 6 96 238 17 167 [86] 308 204 41 6 32 77 14 62 10 3 6 4 2 1 1 1 4 [103] 22 15 13 12 34 14 13 3 1 1 2 2 52 34 6 9 31 [120] 342 348 2 7 52 39 79 5 88 238 40 294 69 878 75 1 6 [137] 5 381 58 84 588 345 161 1293 6 403 516 16 1 1112 54 381 2 [154] 526 38 17 20 17 800 189 1 57 90 92 16 17 31 11 4 17 [171] 12 9 10 46 14 23 1 1 1 313 Can anyone tell me why that could be? Thank you very much, Benjamin
Benjamin Dickgiesser wrote:> Hi, > > I am trying the following: > > library(ismev) > library(evd) > > fit <- gev.fit(x,show=FALSE) > ks.test(x,pgev,fit$mle[1],fit$mle[2],fit$mle[3])1. The test *is* working. It simply warns (and does not report an error) that you have ties in your data and the assumption for calculating p-values do not hold. 2. The KS tests tends to be very conservative given you specify estimated parameters of the distribution. Uwe Ligges> > but I am getting: > Warning message: > cannot compute correct p-values with ties in: ks.test(x, pgev, > fit$mle[1], fit$mle[2], fit$mle[3]) > > where x is: > [1] 239 38 1 43 22 1 5 9 15 6 1 9 156 > 25 3 100 6 > [18] 5 100 10 103 25 5 1 2 11 8 68 13 154 > 67 12 5 15 > [35] 5 130 47 143 176 573 592 213 54 10 17 9 198 > 293 77 11 44 > [52] 6 22 2 10 8 1 2 16 4 70 12 4 7 > 134 41 515 8 > [69] 200 169 2 13 49 218 48 34 74 19 44 128 6 > 96 238 17 167 > [86] 308 204 41 6 32 77 14 62 10 3 6 4 2 > 1 1 1 4 > [103] 22 15 13 12 34 14 13 3 1 1 2 2 52 > 34 6 9 31 > [120] 342 348 2 7 52 39 79 5 88 238 40 294 69 > 878 75 1 6 > [137] 5 381 58 84 588 345 161 1293 6 403 516 16 1 > 1112 54 381 2 > [154] 526 38 17 20 17 800 189 1 57 90 92 16 17 > 31 11 4 17 > [171] 12 9 10 46 14 23 1 1 1 313 > > > Can anyone tell me why that could be? > > Thank you very much, > Benjamin > > ______________________________________________ > R-help at stat.math.ethz.ch mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code.
> cannot compute correct p-values with ties in: ks.test(x, pgev, > fit$mle[1], fit$mle[2], fit$mle[3])You may want to use the ks.boot function in the Matching package which implements a bootstrap ks-test which provides consistent pvalues (achieved significance levels) when there are ties. Cheers, Jas. ======================================Jasjeet S. Sekhon Associate Professor Survey Research Center UC Berkeley http://sekhon.berkeley.edu/ V: 510-642-9974 F: 617-507-5524 ====================================== Benjamin Dickgiesser writes: > Hi, > > I am trying the following: > > library(ismev) > library(evd) > > fit <- gev.fit(x,show=FALSE) > ks.test(x,pgev,fit$mle[1],fit$mle[2],fit$mle[3]) > > > but I am getting: > Warning message: > cannot compute correct p-values with ties in: ks.test(x, pgev, > fit$mle[1], fit$mle[2], fit$mle[3]) > > where x is: > [1] 239 38 1 43 22 1 5 9 15 6 1 9 156 > 25 3 100 6 > [18] 5 100 10 103 25 5 1 2 11 8 68 13 154 > 67 12 5 15 > [35] 5 130 47 143 176 573 592 213 54 10 17 9 198 > 293 77 11 44 > [52] 6 22 2 10 8 1 2 16 4 70 12 4 7 > 134 41 515 8 > [69] 200 169 2 13 49 218 48 34 74 19 44 128 6 > 96 238 17 167 > [86] 308 204 41 6 32 77 14 62 10 3 6 4 2 > 1 1 1 4 > [103] 22 15 13 12 34 14 13 3 1 1 2 2 52 > 34 6 9 31 > [120] 342 348 2 7 52 39 79 5 88 238 40 294 69 > 878 75 1 6 > [137] 5 381 58 84 588 345 161 1293 6 403 516 16 1 > 1112 54 381 2 > [154] 526 38 17 20 17 800 189 1 57 90 92 16 17 > 31 11 4 17 > [171] 12 9 10 46 14 23 1 1 1 313 > > > Can anyone tell me why that could be? > > Thank you very much, > Benjamin > >