Meaning of predictor variable in tutorial

[Piyush Kumar]

The given example on github consists of a response variable y and a predictor x1. Does this mean that 
response time series corresponds to y  and control time series corresponds to x1?



set.seed(1)
x1 <- 100 + arima.sim(model = list(ar = 0.999), n = 100)
y <- 1.2 * x1 + rnorm(100)
y[71:100] <- y[71:100] + 10
data <- cbind(y, x1)

[Kay Brodersen]

Hi Piyush,

Yes, that's exactly it. Note that for all practical purposes, it's strongly recommended to use several control time series (x1, x2, x3, ...).

Kay

--
You received this message because you are subscribed to the Google Groups "CausalImpact" group.
To unsubscribe from this group and stop receiving emails from it, send an email to causalimpact...@googlegroups.com.
To post to this group, send email to causal...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/causalimpact/df27f18c-dc67-4a90-876e-4e3f1bd4f87d%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.