search

facebookresearch / Kats

Kats, a kit to analyze time series data, a lightweight, easy-to-use, generalizable, and extendable framework to perform time series analysis, from understanding the key statistics and characteristics, detecting change points and anomalies, to forecasting future trends.
MIT License
4.93k stars 537 forks source link

!! ValueError: freq T not understood. Please report if you think this is in error. #185

Closed Aditya701 closed 2 years ago

Aditya701 commented 2 years ago

ERROR:root:!! Traceback (most recent call last): !! File "/usr/local/lib/python3.7/dist-packages/kats/detectors/outlier.py", line 114, in detector outlier = self.clean_ts(ts) !! File "/usr/local/lib/python3.7/dist-packages/kats/detectors/outlier.py", line 88, in clean_ts result = seasonal_decompose(original, model=self.decomp) !! File "/usr/local/lib/python3.7/dist-packages/statsmodels/tsa/seasonal.py", line 156, in seasonal_decompose pfreq = freq_to_period(pfreq) !! File "/usr/local/lib/python3.7/dist-packages/statsmodels/tsa/tsatools.py", line 815, in freq_to_period "think this is in error.".format(freq) !! ValueError: freq T not understood. Please report if you think this is in error.

ERROR:root:Outlier Detection Failed

swimmingCreative commented 2 years ago

Hi @Aditya701, can you upload you time series?

Aditya701 commented 2 years ago

ec2_cpu_utilization_5f5533.csv

yangbk560 commented 2 years ago

Hi @Aditya701, I think the problem stems from statsmodels design like mentioned in https://github.com/statsmodels/statsmodels/issues/2856. I would suggest explicitly specifying the seasonality as integer (like 12 for monthly data). Feel free to let us know if this solves your problem. Thanks!