Parallel_PELTSM2: Parallel PELT SM2

In KayleaHaynes/changepoint.parallel: Parallel Changepoint Detection

Description

Splits the data into chunks by sending the first point to the first core, the second point to the second core etc, and runs parallel changepoint detection on each chunk then merges the results and runs parallel detection using the detected changepoints in step 1 as the candidate changepoint locations.

Usage

Parallel_PELTSM2(data, penalty, pruning, sum.stat = norm.sum, cost, minseglen,
  ncores)

Arguments

data	A vector of data-points within which you with to find changepoints.
penalty	The value of the penalty.
pruning	If true PELT is used, if false Optimal Partitioning is used.
sum.stat	This can be “norm.sum" or “exp.sum".
cost	This can be “norm.mean.cost", “norm.var.cost", “norm.meanvar.cost" or “exp.cost".
ncores	Number of cores to use.

Value

The detected changepoints.

Author(s)

Kaylea Haynes

Examples

### Set up parallel environment ###
library(doParallel)
library(foreach)

ncores <- c(10)
cl <- makeCluster(ncores)
registerDoParallel(cl)
### Generate some data from the blocks data set ####
cpts <- round(c(0,0.1, 0.13, 0.15, 0.23, 0.25, 0.40, 0.44, 0.65, 0.76, 0.78, 0.81,1)*10000)
segment_param <- c(0, 4, -1, 2, -2, 3, -1.2, 0.9, 5.2, 2.1, 4.2, 0)
data <- NULL

for (i in 1:(length(cpts)-1)){
  data_new <- rep(segment_param[i],cpts[i+1] - cpts[i])
 data <- c(data, data_new)
}
data <- data + rnorm(10000,0,1)

Parallel_PELTSM2(data, 2*log(length(data)), TRUE, sum.stat = norm.sum, cost = norm.mean.cost, minseglen =1, ncores=10)

KayleaHaynes/changepoint.parallel documentation built on May 7, 2019, 12:29 p.m.