Does FB Prophet generated seasonality assume possible bias by historic periods that are outliers?

[algomaschine]

Hi gents. Imagine we have some really strong moves on the historic data (think BTC 2017 or market moves in October 2008), then

1. when we generate yearly seasonality with FB Prophet, would that year effect the result OR the algorithm is smart enough to remove the outliers?
2. how about monthly seasonality?

Thanks!

[tcuongd]

Prophet does detect trend changes automatically. If there are historical periods you know won't be repeated in your forecast horizon, you can also flag them as one-off holidays to help the model.

I think if the model sees enough data, it'll be able to separate the seasonality from these other effects.

[algomaschine]

Thank you!