@@ahmed007Jaber The goals are very different. Those packages are designed to give good forecasts for real valued time series when domain expertise and external information are limited. mvgam is designed for the opposite (challenging series with missing data, multi series clustering, detection error etc...: nicholasjclark.github.io/physalia-forecasting-course/day1/lecture_1_slidedeck#1)
@@NJ_Clark oh I see. i will have to investigate. what i mentioned came up on my radar as I would like to aggregate daily date into quarterly and monthly time serites and try to forecast until i feel confident enough if data works to forecast daily i am new into all of this time-series analysis and they are on my priority list to learn and experiment with
Thanks for the great introduction. Looking forward to the next in the series!
Ll have to watch this later
Is this better than fable, feast and prophet?
@@ahmed007Jaber The goals are very different. Those packages are designed to give good forecasts for real valued time series when domain expertise and external information are limited. mvgam is designed for the opposite (challenging series with missing data, multi series clustering, detection error etc...: nicholasjclark.github.io/physalia-forecasting-course/day1/lecture_1_slidedeck#1)
@@NJ_Clark oh I see. i will have to investigate. what i mentioned came up on my radar as I would like to aggregate daily date into quarterly and monthly time serites and try to forecast until i feel confident enough if data works to forecast daily
i am new into all of this time-series analysis and they are on my priority list to learn and experiment with