GetWindowLength: Get Window Length

In alaineiturria/otsad: Online Time Series Anomaly Detectors

Get Window Length

Description

GetWindowLength Calculates the size of the window. This window focuses on the real anomaly and it can be used to know if the detected anomaly is a true positive or not.

Usage

GetWindowLength(data.length, num.real.anomaly, window.length.perc = 0.1)

Arguments

data.length	Dataset length.
num.real.anomaly	Number of real anomalies contained in the data set.
window.length.perc	Window length in percentage of the total data

Details

nrow.data and num.real.anomaly must be numeric. Window length is calculated by default as 10% of the length of the data set divided by the number of real anomalies contained in it.

Value

Window length as numeric.

References

A. Lavin and S. Ahmad, “Evaluating Real-time Anomaly Detection Algorithms – the Numenta Anomaly Benchmark,” in 14th International Conference on Machine Learning and Applications (IEEE ICMLA 15), 2015.

Examples

## Generate data
set.seed(100)
n <- 180
x <- sample(1:100, n, replace = TRUE)
x[70:90] <- sample(110:115, 21, replace = TRUE)
x[25] <- 200
x[150] <- 170
df <- data.frame(timestamp = 1:n, value = x)

# Add is.real.anomaly column
df$is.real.anomaly <- 0
df[c(25,80,150), "is.real.anomaly"] <- 1

# Get window length
GetWindowLength(data.length = nrow(df), num.real.anomaly = 3)

alaineiturria/otsad documentation built on Jan. 12, 2023, 12:26 p.m.