KNNdist:

In musto101/wilcox_R: What the package does (short line)

Usage

KNNdist(train, test, g, k = 1, prob = FALSE, plotit = FALSE, xlab = "Group 1", ylab = "Group 2", depthfun = prodepth, ...)

Arguments

train
test
g
k
prob
plotit
xlab
ylab
depthfun
...

Examples

##---- Should be DIRECTLY executable !! ----
##-- ==>  Define data, use random,
##--	or do  help(data=index)  for the standard data sets.

## The function is currently defined as
function (train, test, g, k = 1, prob = FALSE, plotit = FALSE, 
    xlab = "Group 1", ylab = "Group 2", depthfun = prodepth, 
    ...) 
{
    library(class)
    train = as.matrix(train)
    test = as.matrix(test)
    P = ncol(train)
    P1 = P + 1
    xall = as.data.frame(matrix(NA, nrow = nrow(train), ncol = P1))
    xall[, 1:P] = train
    xall[, P1] = g
    xall = elimna(xall)
    x1 = xall[, 1:P]
    xall = as.matrix(xall)
    x1 = as.matrix(x1)
    g = as.vector(xall[, P1])
    ids = unique(g)
    x2 = elimna(test)
    x1 = as.matrix(x1)
    x2 = as.matrix(x2)
    n = nrow(x1)
    n2 = nrow(x2)
    p = length(ids)
    d = matrix(NA, nrow = n, ncol = p)
    D = matrix(NA, nrow = n2, ncol = p)
    for (i in 1:length(ids)) {
        flag = g == ids[i]
        d[, i] = depthfun(as.matrix(x1[flag, ]), pts = x1, ...)
        D[, i] = depthfun(as.matrix(x1[flag, ]), pts = x2, ...)
    }
    res = NULL
    res = knn(d, D, cl = as.factor(g), k = k, prob = prob)
    if (plotit) {
        if (p == 2) {
            plot(d[, 1], d[, 2], xlab = xlab, ylab = ylab, type = "n")
            flag = g == ids[1]
            points(d[flag, 1], d[flag, 2])
            points(d[!flag, 1], d[!flag, 2], pch = "*")
        }
    }
    res
  }

musto101/wilcox_R documentation built on May 23, 2019, 10:52 a.m.