2

According to the documentation the tsfresh algorithm makes use of the Benjamini-Yekutieli in its final step. To give you a little bit of context,

tsfresh automatically calculates a large number of time series characteristics, the so called features. Further the package contains methods to evaluate the explaining power and importance of such characteristics for regression or classification tasks.

I have tried to read the linked references but I found them very technical. Can anybody provide an high-level description of the Benjamini-Yekutieli and explain why it is needed? I would like to understand what is its main purpose. If you don’t know what FRESH is, I would still be happy to read an explanation of the Benjamini-Yekutieli test.

B--rian
  • 5,578
  • 10
  • 38
  • 89
user1315621
  • 3,044
  • 9
  • 42
  • 86
  • Does this [Charles Zaiontz's blog entry](http://www.real-statistics.com/hypothesis-testing/familywise-error/benjamini-hochberg-and-benjamini-yekutieli-tests/) help you? It is still a bit mathematical, but more concise than other resources. – B--rian Dec 17 '19 at 09:31
  • 1
    Edited above. It is an algorithm for extracting a number of features from time series and selecting only the significative ones. – user1315621 Dec 17 '19 at 09:55
  • Thanks for the edit. I just realized, that I do not have library access any more, so I could not check the usual suspects (I mean the first hits in google scholar). May you also add which papers you consulted, please? A [DOI link](https://en.wikipedia.org/wiki/Digital_object_identifier) would be great.. – B--rian Dec 17 '19 at 10:13
  • 1
    OP, this is an interesting topic but suitable for stats.stackexchange.com instead of Stackoverflow. That said, the problem here is that if you're generating lots and lots of features, you have to filter pretty stringently to cut the number of features down to something usable. The B-Y stuff is about making the filter more restrictive the greater the number of features considered. I think the article by Zaiontz referred to above is a good overview. But if you are just using the package, it may not be necessary for you to figure out all of the details. Good luck and have fun. – Robert Dodier Dec 17 '19 at 22:17

0 Answers0