Jeroen Van Goey
2004-Jul-30 15:43 UTC
[R] P-value from the joint cumulative distribution of an n-dimensional order statistic
Hello, I want to compute the P-value from the joint cumulative distribution of an n-dimensional order statistic in R, using the formula found on http://cmgm.stanford.edu/%7Ekimlab/multiplespecies/Supplement/methods_network.html My data consists of three different techniques (G2D, POCUS and RANDOM), and each has associated with it a number of rankings (integer between 0 and 1000), like for example: name r1 r2 r3 r4 r5 r6 r7 r8 r9 r10 1 G2D 939.100 929.400 919.700 910.100 784.200 387.300 377.600 87.130 18.390 17.430 2 POCUS 910.876 910.876 753.944 753.944 0.235 0.208 0.160 0.160 0.016 0.000 3 RANDOM 917.328 893.808 848.294 725.135 462.952 383.682 292.386 46.654 12.466 7.134 I tried already with the package ISMEV from CRAN, (used for order statistics), but the results I get then don't make me much wiser.> library(DBI) > library(RMySQL) > library(ismev) > drv <- dbDriver("MySQL") > con <- dbConnect(drv, group = "mysql") > ibd <- dbSendQuery(con, statement = paste("SELECT * FROM database") ) > ibddata <- fetch(ibd, n = -1) > ibdfit <- rlarg.fit(ibddata[,-1]) >ibdfit$trans [1] FALSE $model $model[[1]] NULL $model[[2]] NULL $model[[3]] NULL $link [1] "c(identity, identity, identity)" $conv [1] 1 $nllh [1] -195.1587 $data r1 r2 r3 r4 r5 r6 1 1.7430e-03 1.8390e-03 3.77600e-03 3.87300e-03 7.84200e-03 8.7130e-03 2 1.0000e-16 8.0000e-15 8.00000e-15 8.00000e-15 1.04000e-14 1.1760e-13 3 7.1338e-02 1.2466e-01 2.92386e-01 3.83682e-01 4.62952e-01 4.6654e-01 r7 r8 r9 r10 1 9.10100e-03 9.19700e-03 9.29400e-03 9.39100e-03 2 3.76972e-11 3.76972e-11 4.55438e-11 4.55438e-11 3 7.25135e-01 8.48294e-01 8.93808e-01 9.17328e-01 $mle [1] 0.1332361 0.5694678 4.2741269 $cov [,1] [,2] [,3] [1,] 3.998775e-12 2.572874e-17 -3.300609e-18 [2,] 2.572874e-17 3.998854e-12 1.264103e-17 [3,] -3.300609e-18 1.264103e-17 3.998948e-12 $se [1] 1.999694e-06 1.999713e-06 1.999737e-06 $vals mu sc xi [1,] 0.1332361 0.5694678 4.274127 [2,] 0.1332361 0.5694678 4.274127 [3,] 0.1332361 0.5694678 4.274127 $r [1] 10 Can ayone explain all these values to me, or point me the way how I can implement the formula from http://cmgm.stanford.edu/%7Ekimlab/multiplespecies/Supplement/methods_network.html in R? Thanks in advance. NB: this message is also crossposted at sci.stat.consult & sci.stat.math