1

I started using R and applying the vars package in R. According to the documentation, there are the options to select none, constant, trend, or both.

I analyzed my time series and were able to plot the graphs, as shown below. However, some feedback on selecting the appropriate VAR model would be great.

The concept of a "trend" (this post is good) is clear, but visually detecting "constant", also known as "drift", or a time series with "non-zero mean", if I'm correct, seems little unclear.

1 enter image description here 2 enter image description here

Community
  • 1
FenleyK
  • 11
  • 2
  • This might be a duplicate of [this thread](http://stats.stackexchange.com/questions/242101/var-models-in-r-constant-trend-constanttrend-none) but I cannot mark it as such as the answer in the other thread has no upvotes yet. Also, is the question more about the options of the `VAR` function in the `vars` package or about time series terminology ("constant", "drift", "trend", etc) in general? – Richard Hardy Oct 28 '16 at 17:43

0 Answers0