A Time Series is Worth 64 Words: Long-term Forecasting with Transformers
β AK (@_akhaliq) November 29, 2022
abs: https://t.co/vEvAaM5KfE pic.twitter.com/SHH25hIVyc
A Time Series is Worth 64 Words: Long-term Forecasting with Transformers
β AK (@_akhaliq) November 29, 2022
abs: https://t.co/vEvAaM5KfE pic.twitter.com/SHH25hIVyc
Spoke to a data scientist who was frustrated that his company decided to buy a proprietary demand forecasting system when the performance of his XGBoost model was so much stronger.
β Anthony Goldbloom (@antgoldbloom) July 8, 2022
His solution: he fed his XGBoost forecasts into the proprietary system as a feature π
Forecasting Future World Events with Neural Networks
β AK (@_akhaliq) July 1, 2022
abs: https://t.co/tD8F0ZC1rC
github: https://t.co/v8HZgye0ZH
a dataset for measuring the ability of neural networks to forecast future world events, containing thousands of forecasting questions and an accompanying news corpus pic.twitter.com/xsQnxfgdia
Our 50+-year review of #forecast combinations is now out. @YanfeiKang @f3ngli @Xia0qianWang https://t.co/MTBhI2trNe
β Rob J Hyndman (@robjhyndman) May 10, 2022
This looks really exciting for time series work. The original project from @robjhyndman in R is an absolute classic, and to have it available in Python is huge.
β Jeremy Howard (@jeremyphoward) February 24, 2022
Even better - it's an #nbdev project which means it's got great docs and tests and is easy to contribute to. https://t.co/22o6D7EHFQ
Monash time series forecasting repository https://t.co/Mcc6N4fiJL #rstats #forecasting
β Rob J Hyndman (@robjhyndman) February 23, 2022
CoST: Contrastive Learning of Disentangled Seasonal-Trend Representations for Time Series Forecasting
β AK (@ak92501) February 4, 2022
abs: https://t.co/4jk8lps3pJ pic.twitter.com/WJmZRWrYMk
ETSformer: Exponential Smoothing Transformers for Time-series Forecasting
β AK (@ak92501) February 4, 2022
abs: https://t.co/ZtpXPqhlhF pic.twitter.com/dSPGXgcAid
Want to know why it's risky to assume that your ML is going to continue work if the test regime changes from your training data? Just ask Zillow.
β Gary Marcus (@GaryMarcus) November 12, 2021
https://t.co/EUH71gZspy
The thing Iβm stuck on about Zillow Data Science Discourse is the data scientistsβ models could have been far more accurate as humans at predicting housing prices and still have gotten this outcome.
β Jacqueline Nolis (@skyetetra) November 5, 2021
Merlion - A Machine Learning Framework for Time Series Intelligence. https://t.co/j6vYdpg3xZ #Python #MachineLearning pic.twitter.com/gw9ng0yiCm
β Python Weekly (@PythonWeekly) September 28, 2021
"Modern Time Series Analysis with STUMPY" -- really enjoyed this SciPy2021 talk by @seanmylaw. Great intro to subsequence comparison-based time series analysis and using matrix profiles. Really makes me want to find a time series problem to work on new π€ https://t.co/KcWn1Zjt0D pic.twitter.com/QHTuzBULKR
β Sebastian Raschka (@rasbt) September 6, 2021