Can Not Rolling With Binary Feature

[b-y-f]

The problem:

import pandas as pd
df = pd.DataFrame({
   "id": [1, 1, 1, 1, 2, 2],
   "time": [1, 2, 3, 4, 8, 9],
   "x": [1, 2, 3, 4, 10, 11],
   "y": [True, True, True, False, False, False],
})
from tsfresh.utilities.dataframe_functions import roll_time_series
df_rolled = roll_time_series(df, column_id="id", column_sort="time")
from tsfresh import extract_features
df_features = extract_features(df_rolled, column_id="id", column_sort="time")
df_rolled

Got error

Anything else we need to know?:

File /opt/homebrew/Caskroom/miniforge/base/envs/lake/lib/python3.10/site-packages/tsfresh/feature_extraction/feature_calculators.py:315, in symmetry_looking()
    313     x = np.asarray(x)
    314 mean_median_difference = np.abs(np.mean(x) - np.median(x))
--> 315 max_min_difference = np.max(x) - np.min(x)
    316 return [
    317     ("r_{}".format(r["r"]), mean_median_difference < (r["r"] * max_min_difference))
    318     for r in param
    319 ]

TypeError: numpy boolean subtract, the `-` operator, is not supported, use the bitwise_xor, the `^` operator, or the logical_xor function instead.

Environment:

• Python version:3.10
• Operating System:MAC
• tsfresh version: latest
• Install method (conda, pip, source):

[nils-braun]

Actually, it is not the rolling which is problematic but the feature extraction after that. The feature calculator inside of tsfresh do only work for numerical features. For binary features most of them are not defined well. What you can do is convert your boolean into -1 and 1 or 0 and 1 and then use the feature extraction.

[b-y-f]

@nils-braun Thanks, changed to 0,1