blue-yonder / tsfresh

Automatic extraction of relevant features from time series:

Home Page:http://tsfresh.readthedocs.io

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

If I have a series data for only one id and I don't want extract feature and compress to one row?

b-y-f opened this issue · comments

commented

The problem:

If I have a series data for only one id and I don't want extract feature and compress to one row? I just want to generate more column like lagged f1 lagged f2, f1 or f2 relation to the previous or past time?

ts     f1    f2  label
1       a     b    True
2      c     d    True
3      e     f    False
4      c     g    True
5      y     r    True
......

I want:

ts    f1    f2    added_feature 1  label
1      a      b    A True
2     c      d    B  True
3     e      f    C  False
4     c     g    D    True
5     y      r     E  True
......

Anything else we need to know?:

Environment:

  • Python version:
  • Operating System:
  • tsfresh version:
  • Install method (conda, pip, source):

Hi @b-y-f, the solution is to create rolling windows and extract features for each of the windows.

We have documentation available on this topic here: https://tsfresh.readthedocs.io/en/latest/text/forecasting.html