SAYAN CHAKRABORTY

Migrate bayesian-changepoint-detection dependency to bayescd. reference: https://github.com/hildensia/bayesian_changepoint_detection/pull/35#issuecomment-1089303593

Related to: https://github.com/py4j/py4j/issues/476

Change the logic for model expiry to optional based on model freshness

2

The current logic expires the model after k time units depending on the data frequency. This prevents forecasting for long terms. The expiry logic should be change to optional.

_detect_window_size is currently not optimized for weekly time series data in order to detect the most frequent periodic pattern. This issue need some investigation on that front. Reference: https://github.com/zillow/luminaire/pull/114 Note:...

Reference: https://github.com/papaemman/luminaire_demo TL:DR; Data Exploration module doesn't work for time series with frequency 'W' because of a missing variable (tc_window_length). Any idea why is this happening?

Related issue for the batch module as reference: https://github.com/zillow/luminaire/pull/125