simulate: A function for generating simulated multivariate data

Description

Generates multivariate simulated data having n observations and p variates. The data have a standard Gaussian distribution except at a specified number of locations where there is a change in mean in a proportion of the variates. The function is useful for generating data to demonstrate and assess multivariate anomaly detection methods such as capa.mv and pass.

Usage

simulate(n = 100, p = 10, mu = 1, locations = 40, durations = 20,
  proportions = 0.1)

Arguments

The number of observations. The default is n=100.

The number of variates. The default is p=10.

The change in mean. Default is mu=1.

locations

A vector of locations (or scalar for a single location) where the change in mean occurs. The default is locations=20.

durations

A scalar or vector (the same length as locations) of values indicating the duration for the change in mean. If the durations are all of the same length then a scalar value can be used. The default is durations=20.

proportions

A scalar or vector (the same length as locations) of values in the range (0,1] indicating the proportion of variates at each location that are affected by the change in mean. If the proportions are all same than a scalar value can be used. The default is proportions=0.1.

Value

A matrix with n rows and p columns

Examples

Run this code

# NOT RUN {
library(anomaly)
sim.data<-simulate(500,200,2,c(100,200,300),6,c(0.04,0.06,0.08))

# }

Run the code above in your browser using DataCamp Workspace