find_HDoutliers: Detect Anomalies in High Dimensional Data.

In stray: Anomaly Detection in High Dimensional and Temporal Data

Description

Detect anomalies in high dimensional data. This is a modification of HDoutliers.

Usage

find_HDoutliers(
  data,
  alpha = 0.01,
  k = 10,
  knnsearchtype = "brute",
  normalize = "unitize",
  p = 0.5,
  tn = 50
)

Arguments

data	A vector, matrix, or data frame consisting of numerical variables.
alpha	Threshold for determining the cutoff for outliers. Observations are considered outliers if they fall in the (1- alpha) tail of the distribution of the nearest-neighbor distances between exemplars.
k	Number of neighbours considered.
knnsearchtype	A character vector indicating the search type for k- nearest-neighbors.
normalize	Method to normalize the columns of the data. This prevents variables with large variances having disproportional influence on Euclidean distances. Two options are available "standardize" or "unitize". Default is set to "unitize"
p	Proportion of possible candidates for outliers. This defines the starting point for the bottom up searching algorithm. Default is set to 0.5.
tn	Sample size to calculate an emperical threshold. Default is set to 50.

Value

The indexes of the observations determined to be outliers.

References

Wilkinson, L. (2018), 'Visualizing big data outliers through distributed aggregation', IEEE transactions on visualization and computer graphics 24(1), 256-266.

Examples

require(ggplot2)
set.seed(1234)
data <- c(rnorm(1000, mean = -6), 0, rnorm(1000, mean = 6))
outliers <- find_HDoutliers(data, knnsearchtype = "kd_tree")



set.seed(1234)
n <- 1000 # number of observations
nout <- 10 # number of outliers
typical_data <- matrix(rnorm(2 * n), ncol = 2, byrow = TRUE)
out <- matrix(5 * runif(2 * nout, min = -5, max = 5), ncol = 2, byrow = TRUE)
data <- rbind(out, typical_data)
outliers <- find_HDoutliers(data, knnsearchtype = "brute")

stray documentation built on July 2, 2020, 4:03 a.m.