Given a time series, suppose I wish to determine if trend is present. I understand I can use the Cox-Stuart test (https://www.r-bloggers.com/trend-analysis-with-the-cox-stuart-test-in-r/), but I have also read that the Augmented Dickey Fuller test can be done for this (https://machinelearningmastery.com/time-series-data-stationary-python/). More specifically, the last link says:

"The Augmented Dickey-Fuller test is a type of statistical test called a unit root test.The intuition behind a unit root test is that it determines how strongly a time series is defined by a trend."

But I have also read from (https://kourentzes.com/forecasting/2017/03/30/can-you-spot-a-trend-in-a-time-series/):

"I do not think it is an exaggeration to suggest that even experts often can be confused on the exact definition (and effect) of a unit root and a trend."

So my questions are thus:

  1. What does it mean for a time series to have a unit root? Does that automatically mean there is a trend in the time series?
  2. What is the difference between the Cox-Stuart and Augmented Dickey Fuller Tests? Which should I definitely use for trend detection?

There are two types of trend: stochastic and deterministic. Unit roots are a type of stochastic trend that the augmented Dickey-Fuller tests for. The Cox-Stuart test is for deterministic trends.

| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.