Linear Decomposition Time Series Model (TS E5)

preview_player
Показать описание
In the previous video we created a multiplicative model but now we want to see if we can create a linear (additive) model. Using the decomposition of a time series into trend and cycle we will create a model that is based on adding the trend and cycle together to get a prediction. Seasonality and sub-seasonality could be added however for this example we will keep it simple.
Рекомендации по теме
Комментарии
Автор

Videos like this is why I am subscribed to your channel!

davidhammar
Автор

Thank you for these videos. Learning a lot!

hacker
Автор

Do you plan to make a video on the effectiveness of time series (or generally models) to predict the future based on data collected in the past? This is basically the core aspect of Nassim Talebs Black Swan Theory. I think a video on that (or his other works in the Incerto series) would be really interesting for those unfamiliar with the idea and for the rest to hear your opinion on it.

dfa
Автор

I have a question regarding the cyclical component:
Is there any mathematical way to derive the frequency of the cycle directly from the data without guessing? It really seems like there should be as the cyclical change of whatever variable influences your y is of course reflected in the data.

dfa
Автор

Hey Dimitri is it possible for someone that studied actuarial science to movie into the world of becoming a quant

marcusbirdman