initAFFMeanCD: Initialisation of AFF change detector

In ffstream: Forgetting Factor Methods for Change Detection in Streaming Data

Initialisation of AFF change detector

Description

This function makes it simple to initalise an FFF object.

Usage

initAFFMeanCD(alpha = 0.01, eta = 0.01, BL = 50)

Arguments

alpha	The value of the significance level. Default value is 0.01, although it is recommended that the user set this parameter.
eta	The value of the step-size in the gradient descent. Default is eta=0.01.
BL	The length of the burn-in region. Default value is BL=50. Must be at least greater than or equal to 2. No maximum. However, there is an exception: BL=0 also works, but in this case the user needs to specify the streamEstMean and streamEstSigma; see

Value

An AFF mean change detector object.

Examples

library(Rcpp)
affmeancd1 <- initAFFMeanCD()              # initialises with alpha=0.01, 
                                           # eta=0.01 and BL=50

affmeancd2 <- initAFFMeanCD(alpha=0.005, eta=0.1,  BL=100) 


affmeancd3 <- initAFFMeanCD(alpha=0.005, eta=0.1,  BL=0)     # Example 3
affmeancd3$streamEstMean <- 0
affmeancd3$streamEstSigma <- 1

ffstream documentation built on May 31, 2023, 7:53 p.m.