r - distribution comparison for large number of data sets -

-

i have data named "data" this:

centre_blinded  val_list 1   1104    c(-13, -1, 0, 28, -88, 28, -1, -6, -5, -58, 28, 28, 28, 28, 2, 0, 28, -26, 28, 28, 2, 28,    28, -2, -29, 0, 28, -34, -6, 0, 28, 1, 0, 0, -1, 28, 28, 0, 28, 6, 28, 0, 28, 28, 28, 0, -2, -6, -1, 4, 6, 1, -16, -7, 2, 3, 7, 0, 1, 11, 0, 1, -6, -5, 0, 3, 8, 7, 0, 0, 6, -6, 2, 36, -8, 0, -7, -7, -1, -1, -1, 7, -3, 7, 2) 2   1204    c(2, -9, 28, 28, -2, 1, -3, -1, 0, 28, 28, 28, 28, 28, 28, 3, 10, -5, -8, 9, -8, 0, 13, 0, -1, 2, -1, 0, 6, 1, 0, -7, 6, -6, 1) 3   1403    c(0, 2, 0, 2, 28, 0, -1, -35, -36, 2, 1, 1, 28, 28, 28, 0, 0, 28, -7, -35, 28, -3, -18, 28, 28, 28, -5, 0, 28, -2, 4, 5, 0, 56, 1, 0, 1, -7, -20, 0, 0, -3, 0, 1, 3, 0, 4, -2, 42, -13, 7, 10, 7, 56, 0, -5, 10, 56, 8, 56, 84, -4, 1, 0, -14, -7, -1, -48, -6, -3, 0, 7) 4   1110    c(0, 1, 0, -3, 28, 28, 0, -5, 0, 9, 15, 56, -11, -1, -7)

the first column containts id of centre, , second contains list of values. wanto build empirical distributions these values each centre , compare them pairwaise using e.g. kolmogorov-smirnov test (ks.test in r). way n x n matrix of p-values. question how , preserve ids of centres each k-s test.

my try was:

val_list_temp = as.list(data, = "centre_blinded")) val_list = val_list_temp[[2]] names(val_list) = val_list_temp[[1]]

here have ids each centre, when use expand.grid don't no how store them more:

val_table = as.data.table(expand.grid(val_list, val_list)) ks_tests = apply(x = val_table, 1, function(x) ks.test(unlist(x[1]),unlist(x[2])))

besides, how put in matrix later, visualisation?

p.s. maybe there better way using kolmogorov-smirnov test , making matrix of p-values?

is type of solution looking for? loops 6 times (4 choose 2) , makes matrix symmetrical.

#data provided a1104 <- c(-13, -1, 0, 28, -88, 28, -1, -6, -5, -58, 28, 28, 28, 28, 2, 0, 28, -26, 28, 28, 2,         28,28, -2, -29, 0, 28, -34, -6, 0, 28, 1, 0, 0, -1, 28, 28, 0, 28, 6, 28, 0, 28,         28, 28, 0, -2, -6, -1, 4, 6, 1, -16, -7, 2, 3, 7, 0, 1, 11, 0, 1, -6, -5, 0, 3, 8, 7,         0, 0, 6, -6, 2, 36, -8, 0, -7, -7, -1, -1, -1, 7, -3, 7, 2) a1204 <- c(2, -9, 28, 28, -2, 1, -3, -1, 0, 28, 28, 28, 28, 28, 28, 3, 10, -5, -8, 9, -8, 0, 13,        0, -1, 2, -1, 0, 6, 1, 0, -7, 6, -6, 1) a1403 <- c(0, 2, 0, 2, 28, 0, -1, -35, -36, 2, 1, 1, 28, 28, 28, 0, 0, 28, -7, -35, 28, -3, -18,         28, 28, 28, -5, 0, 28, -2, 4, 5, 0, 56, 1, 0, 1, -7, -20, 0, 0, -3, 0, 1, 3, 0, 4, -2,         42, -13, 7, 10, 7, 56, 0, -5, 10, 56, 8, 56, 84, -4, 1, 0, -14, -7, -1, -48, -6, -3, 0, 7) a1110 <- c(0, 1, 0, -3, 28, 28, 0, -5, 0, 9, 15, 56, -11, -1, -7)  data <- list(a1104,a1204,a1403,a1110) names(data) <- c("a1104","a1204","a1403","a1110")  #provide combinations want compare varcomb = combn(1:length(data), 2)  #create matrix of zeros , populate ks.test result.matrix <- matrix(0,nrow=length(data),ncol=length(data)) for(i in 1:(dim(varcomb)[2])){    result.matrix[varcomb[1,i],varcomb[2,i]] <- ks.test(data[[varcomb[1,i]]],data[[varcomb[2,i]]])$p.value }  result.matrix <- result.matrix+t(result.matrix) row.names(result.matrix) <- names(data) colnames(result.matrix) <- names(data) result.matrix

Comments

Post a Comment

Popular posts from this blog

windows - Single EXE to Install Python Standalone Executable for Easy Distribution -

-
i used pyinstaller create standalone exe file pyqt project. did not use single file option because exe created single file option takes long open. therefore, pyinstaller created dist folder contains bunch of files including program.exe file. create setup.exe file place contents of dist folder in program directory , create shortcut in start menu , on desktop. want super simple user. maybe setup.exe files when download program cnet. found inno-setup, looks promising. however, not know if there special considerations because program standalone program create python/pyqt program. thanks! have experience task? there program using task not know about? inno-setup or nsis easiest use. tell them files include , put them , create nice installer you. wrote short tutorial on experiences using innosetup might find helpful: http://www.blog.pythonlibrary.org/2008/08/27/packaging-wxpymail-for-distribution/ note tutorial based around wxpython app wrote, concepts same.
Read more

c# - Access objects in UserControl from MainWindow in WPF -

-
i have simple xaml, how can change text property of textblock in usercontrol1 mainwindow ? <window x:class="refactorxaml.mainwindow" xmlns="http://schemas.microsoft.com/winfx/2006/xaml/presentation" xmlns:x="http://schemas.microsoft.com/winfx/2006/xaml" xmlns:refactorxaml="clr-namespace:refactorxaml" title="mainwindow" height="350" width="525"> <grid> <grid.rowdefinitions> <rowdefinition height="100"></rowdefinition> <rowdefinition height="100"></rowdefinition> </grid.rowdefinitions> <grid> <refactorxaml:usercontrol1></refactorxaml:usercontrol1> </grid> <grid grid.row="1"> <button>change text</button> </grid> </grid> </window> <usercon...
Read more

javascript - How to name a jQuery function to make a browser's back button work? -

-
i'm trying write function ajaxyfy web site, works perfect, except function button. that's have now: function loading() { if (typeof history.pushstate !== "undefined") { var historycount = 0; $('.menu-item, .logo').on('click','a',function(e){ e.preventdefault(); $(this).data('clicked', true); if ($(this).is('.logo a')) { var homepage = $(this).attr('href'); function1(homepage); history.pushstate(null, null, homepage); } else if ($(this).is('.projects a')) { var projects = $(this).attr('href'); function2(projects); history.pushstate(null, null, projects); } else { var pages = $(this).attr('href'); function3(pages); history.pushstate(null, null, pages); ...
Read more