Description Usage Arguments Examples
View source: R/detect_outliers.R
Detects outliers in a distance matrix using a certain method and following a certain criterion.
1 2 | detect_outliers(distMatrix, method = "MD", criterion = "MAD", LOF_k = 2,
MAD_trim = 2, boxplot_trim = 1.5)
|
distMatrix |
Numeric, distance matrix |
method |
Character, method for measuring separation of one point
from all other points.
The following are accepted:
"MdD": median distance;
"MD": average (mean) distance;
"MAH": Mahalanobis distances ( |
criterion |
Numeric/Character, the criterion used for separating outliers. The following are accepted: <Numeric, 0-1>: number between 0 and 1, sets a quantile type of threshold; "boxplot": outliers are those singled out as points in a boxplot; "MAD": threshold is given by Median Absolute Deviation. |
LOF_k |
Numeric, when method = "LOF", the size of the neighborhood.
See |
MAD_trim |
Numeric, when criterion = "MAD", the multipler of MAD to calculate a outlier threshold. |
boxplot_trim |
Numeric, when criterion = "boxplot", the multipler of the interquartile range (IQR) to calculate a outlier threshold. |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 | ## Not run:
pca <- princomp(iris[, 1:4])
irisOutliers_MD <- detect_outliers(dist(iris[, 1:4]),
method = "MD",
criterion = "MAD")
irisOutliers_LOF <- detect_outliers(dist(iris[, 1:4]),
method = "LOF",
criterion = "MAD")
plot(pca$scores[, 1:2], col = "black", main = "Outliers")
points(pca$scores[irisOutliers_MD$index, 1:2],
col = "red", pch = 2, cex = 1.5)
points(pca$scores[irisOutliers_LOF$index, 1:2],
col = "blue", pch = 6, cex = 1.5)
legend(0.65 * max(pca$scores[,1]), max(pca$scores[,2]),
c("MD", "LOF"), pch = c(2, 6), col = c("red", "blue"))
## End(Not run)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.