Change Detection

library(cpm)

Change Detection

1. Detect if there was a change
2. Estimate change point

Change in mean, median

# simulate data
set.seed(0)
x <- c(rnorm(200, 0, 1), rnorm(200, 0.5, 1))

resultsMW <- detectChangePointBatch(x, cpmType = "Mann-Whitney", alpha = 0.05)
par(mfrow = c(1, 2))
plot(x, type = "l", xlab = "time", ylab = "x", bty = "l")
if (resultsMW$changeDetected)
  abline(v = resultsMW$changePoint, lty = 2, col = "red")

plot(resultsMW$Ds, type = "l", xlab = "time", ylab = expression(D[t]), bty = "l")
abline(h = resultsMW$threshold, lty = 2)