A function to calculate all components of Ueda's method.

find_ueda_outliers(x, smax = 5)

Arguments

x

A numeric vector

smax

The maximum number of outliers to be detected. By default, smax = 3.

Value

A list of 4 components:

1. The Ut matrix.

2. The label matrix.

3. The input numeric vector.

4. The input vector with no outliers.

References

Marmolejo-Ramos, F., Vélez, J.I. & Romão, X. Automatic detection of discordant outliers via the Ueda’s method. J Stat Distrib App 2, 8 (2015). https://doi.org/10.1186/s40488-015-0031-y

Examples

# random seed for reproducibility
set.seed(13)
# introduce 5 outliers
x <- c(rnorm(25, 300, 10), rnorm(5, 400, 5))
# shuffle the data
x <- sample(x)
# removes up to 5 outliers
find_ueda_outliers(x, smax = 5)
#> $Ut
#>        none    X30    X29    X28    X27     X26
#> none -0.509  0.452  0.888  0.113 -3.256 -17.821
#> X1    2.946  3.896  4.323  3.542  0.122 -15.879
#> X2    6.411  7.358  7.789  7.028  3.638 -13.298
#> X3    9.819 10.760 11.197 10.466  7.137 -10.555
#> X4   13.106 14.038 14.475 13.766 10.489  -8.291
#> X5   16.398 17.323 17.766 17.101 13.950  -4.802
#> 
#> $label
#>        none    X30    X29    X28    X27      X26
#> none -0.509  0.452  0.888  0.113 -3.256 -17.821*
#> X1    2.946  3.896  4.323  3.542  0.122  -15.879
#> X2    6.411  7.358  7.789  7.028  3.638  -13.298
#> X3    9.819  10.76 11.197 10.466  7.137  -10.555
#> X4   13.106 14.038 14.475 13.766 10.489   -8.291
#> X5   16.398 17.323 17.766 17.101  13.95   -4.802
#> 
#> $x
#>  [1] 281.4397 286.3902 289.0641 289.5459 295.6014 296.3462 297.1973 298.0605
#>  [9] 298.8556 301.0066 301.8732 302.3668 302.6254 303.5740 304.1553 304.6187
#> [17] 305.5433 306.2018 307.0223 311.0514 311.4253 312.2951 313.9643 317.7516
#> [25] 318.3616 389.8648 392.7034 394.7152 396.3593 400.7468
#> 
#> $x_new
#>  [1] 281.4397 286.3902 289.0641 289.5459 295.6014 296.3462 297.1973 298.0605
#>  [9] 298.8556 301.0066 301.8732 302.3668 302.6254 303.5740 304.1553 304.6187
#> [17] 305.5433 306.2018 307.0223 311.0514 311.4253 312.2951 313.9643 317.7516
#> [25] 318.3616
#>