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
4.85k
stars
530
forks
source link
Implementation of online/real time inference for detection of change point by bocpd #323
It seems the bocpd's detector has not implemented an online/real time inference which means, after fitting the underlying model for some historical values and when there is a novel data point observed, the detector does not accept a new data point and make inference based on this new data point. The implementation of _BayesOnlineChangePoint._find_posterior() assumes there is always a whole series submitted and it needs to start from step 1 initialization.
Please let me know if I have misunderstanding. It would be grateful if an example can be provided to demonstrate the online inference.
Thank you
It seems the bocpd's detector has not implemented an online/real time inference which means, after fitting the underlying model for some historical values and when there is a novel data point observed, the detector does not accept a new data point and make inference based on this new data point. The implementation of _BayesOnlineChangePoint._find_posterior() assumes there is always a whole series submitted and it needs to start from step 1 initialization. Please let me know if I have misunderstanding. It would be grateful if an example can be provided to demonstrate the online inference. Thank you