What is the Prophet Model

[@baharl2981]

wow, you can't begin to understand how easy you explained everything I needed to know THANK YOU !

[@EngineeringChampion]

Aric, I've watched all your videos in this series. I love the fact that you talk fast and move on. You give a chance to my brain to think about another chess game. Thank you!

[@alisavictory2969]

Great and concise! Very engaging especially with the witty titles! Thank you for sharing :)

[@alexei.domorev]

Brilliant and concise explanation! Aric, please keep them coming. ;)

[@MyMy-tv7fd]

strangely good, I wish all model explanations were like this

[@zedor1553]

strangely good, i feel you

[@jacobmoore8734]

Fourier: "for" - "ee" - "ay"

[@pipertripp]

Not to be confused with Furries, which is a completely different beast.

[@rokaskarabevicius]

Excellent videos, can't wait for more.

[@wesamalsohle5384]

could you give me the link for your presentation please, thanks for your time.

[@AricLaBarr]

Sorry, but there is no link for the presentation!

[@josephtolentino1900]

Short but concise 😊

[@pipertripp]

I like the Fourier series idea. That is clever.

[@TheBlackNight971]

I think there is an error. The "Term" in the fourier series is considered as a pair sin(x)+cos(x) for each n. So if we choose n = 3 (number of terms) we will have 3 pairs of sin(x)+cos(x). @Aric LaBarr

[@MaxGroßeHerzbruch]

so is it possible to read out the algebraic form of the fitted linear model out explicitly?:) If not, is there another approach where this is possible?

[@AricLaBarr]

It definitely is possible! The more complicated the model (more knots, more fourier terms, more holidays) the more complicated the equation is. y = beta0 + TREND + FOURIER + HOLIDAY TREND = piecewise linear regression on trend. That could be a simple as beta1*time (no knots) or more complicated with knots FOURIER = beta term multiplied by each of the fourier terms described in the video HOLIDAY = beta term multiplied by dummy variable where it is a 1 for the holiday and 0 otherwise

[@hogrideeeeer]

Hi Aric, could you make a video on VAR modelling?

[@AricLaBarr]

That is definitely on the upcoming list!

[@happyoblap]

Holy fuck...what an amazing video

[@anoriginalnick]

It sounds like a fancy wrapper over Python's time series decomposition with structural breaks embdedded. It does feel very overfit. Any thoughts on in this ?

[@AricLaBarr]

There are definitely some similarities with time series decomposition. Most of the time, in TS decomposition we use LOESS to estimate the underlying trend as compared to just piecewise regression. Seasonality it also estimated differently, but you are correct in the idea that they handle the pieces individually!

[@levi2732]

whait we use trigonometry in inferential statistics? that's awesome

[@hahahNo197]

Lesseon algoritma prophet

[@reyes09071962]

Four-YAE

[@diysumit]

111th viewer