I am very new to this data analytics field, so bit confused about what to do. I have 2 time series datasets of two different products that have the same attributes. I want to find trend similarity between both the time series. Which one will be best to use, correlation, covariance, mutual information or Euclidean distance?
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
2
what do you mean by "trend similarity". if the data has a trend assuming it's linear you can run the regression of data on the index and then judge if the coefficients are significant or not. but this won't give any insight on similarity of the trends; maybe you can compare the slopes for equality.
you cannot really apply correlation or covariance because I do not think your data is stationary in presence of a trend
-
$\begingroup$ by trend similarity I mean the similarity of the trend of both the series over time, like one is moving down then another one is also moving down and vice versa. $\endgroup$ Commented Jun 9, 2023 at 9:44
-
$\begingroup$ as i said you can regress on the "index" (an increasing sequence from 1 to n) and then judge the slope based on t stats. $\endgroup$– pppp_prsCommented Jun 9, 2023 at 17:49