Max Triba

Energy sector data science ⚡️

[kelly victoria 🪐 @saysthefox]

Our book is finally here! Sign up to get your copy! There are tons of extra special add-on items and limited edition collectibles. Top 5 backers will get to add their own character to the book! It may have taken 6 years but it was worth the wait. 🧚‍♂️

https://t.co/NT5TP21ij3

[pikuma.com @pikuma]

Sum of consecutive odd integers:

[Max Triba @maxtriba]

@Jac5Connor Liquidation for margin calls on nvda, etc?

[Max Triba @maxtriba]

@Jac5Connor Ran into a classmate from college (Nebraska - 7 years prior) randomly while giving my sister and BIL a tour of Omotesando.

[Max Triba @maxtriba]

@DylanMcD8 I still miss the G4 PowerBook keyboards

[Max Triba @maxtriba]

@tony_gazillioni @__paleologo 0.38 is where it’s at

[Max Triba @maxtriba]

@JohnPhamous @basementstudio I’d love to learn how to make this

[Max Triba @maxtriba]

@seandsweeney Not a developer (yet!) but my inspiration collection is coming along

[Max Triba @maxtriba]

@JohnHelveston Still here 🥲

[Max Triba @maxtriba]

Had a dream my cat could see into the future, but only 10 minutes. He wanted to use it to trade FX.

[Max Triba @maxtriba]

@DaveMcNamee3000 @7eleven please send this man some marby lights and a slim jim

[Max Triba @maxtriba]

@arithmoquine https://t.co/CCrGNSykz9

[Max Triba @maxtriba]

@quantymacro @0xfdf @quant_arb Luckily with some time on Paul’s Math Notes (basic algebra through diff eq, with solutions) (https://t.co/kqNvlc0JER) and Gil Strang’s MIT linear algebra (https://t.co/UQoDExZNTW) + 3b1b, we’ll get there.

[jack @jack24dd30]

just got this text from a trusted friend in DC

[Valeriy M., PhD, MBA, CQF @predict_addict]

One of the early time series forecasting competition was held in 1991 by physicists at Santa Fe institute and by this day it remains the most serious time series competition that has resulted in 630+ pages book.

Competition included 6 rather hard datasets from diverse subjects:

- physics laboratory experiment describing fluctuations in a far-infrared laser

- physiological data from a patient with sleep apnea (heart rate, chest volume, blood oxygen and EEG sleeping state)

- high frequency currency exchange data

- synthetic series designed for the competition

- astrophysical data from a variable star

- a fugue. Bach’s unfinished fugue from The Art of Fugue.

Compared to M competitions that mostly employed very short time series (less than 100 values) and focused on primitive linear models in Santa Fe all the successful solutions were fundamentally non linear and required much more expertise and manual control.

The competition was won by machine learning neural network and state space reconstruction method.

The Santa Fe competition also asked participants to include error bars to indicate uncertainty around the forecasts that M competitions didn’t have … until 2018, almost 30 years after the Santa Fe competition had probabilistic forecasts.

Some Santa Fe dataset examples in comments.

#timeseries #forecasting

[Max Triba @maxtriba]

@bnj @linaeons Close to another city lmao

[Max Triba @maxtriba]

@juanbuis Cat has the same thousand-yard-stare mine does when I’m scooping his doodoo

[i know now @powcampsurvivor]

i’m so obsessed with this chinese influencer who goes through like a million steps just to make little treats that never look that good