Giters

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.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

What parameters can be used for specify change directions in bayesian and robust changepoint detection models?

jscanass opened this issue · comments

commented

Hi there! In the detection tutorial notebook it is shown for the CUSUMDetector model a parameter, change_directions, for specifying what type of change to be detected. What is the analogous parameter in the case of BOCPDetector, and RobustStatDetector? Or in which other form I can specify an increase or decrease? A working example could be the same as the tutorial.

commented

@jscanass There is not a straightforward way to pre-specify the change direction in either of these detectors. For these detectors you will need to calculate all change points and then filter afterwards if you wish.