@KabirCreates As a follow-up, I showed this parlor trick to my stats class. I am covering regression these past few weeks. Perfect timing to show this case and the associated ethical issues.
@KabirCreates You are spot on. Just smoke and mirrors with y and x variables to give a negative scatter to make the gullible think there is a red flag. At the core, he is assuming Trump should receive the same % of party voters as candidate voters. That is a red herring and false assumption.
@UChiPritzker CNN today mentions Scott Atlas as an alumus of uchicago med as do other media. Stanford Med wrote an open letter disavowing him. Will uchicago take a stance?
Pioneering #UChicago mathematician, Prof. Gregory Lawler, has been awarded the prestigious Wolf Prize for his work in probability.
Join us in congratulating Prof. Lawler on receiving this honor! https://t.co/XzXnOvJV5l
@alwaysmansour Dear Mansour, I have always been proud to carry my name. I realized I should do it before Twitter became popular. Thxs for kindly asking. I love you brand name!
My university @UChicago continues to be an Economics giant. Paul Romer, an alumnus, wins share of Nobel Prize. Congrats!
https://t.co/kdSpd0FjTd
91 scholars associated with the university have received Nobel Prizes so far. 30 of them in economics. @UChi_Economics
@spyrosmakrid Great ML/time series paper. I am making it required reading for my students. Statisicians need to do more studies like this. ML can have a role but folks accept it like a cult without studies like yours.
ML methods can't even beat out naive forcast. Statisticians need to do more studies like this to show at times the Emperor (ML) has no clothes. https://t.co/KfHbwOsMBp
The major finding is: ”The forecasting accuracy of ML forecasting methods is lower than the worst of statistical ones while the accuracy of more than half the ML methods is lower than a random walk”.
https://t.co/6aMDaHKrnH
@brianwilt@Jawbone Nice distribution plot of mean sleep times by University on jawbone blog. What is the mean and standard deviation of individual student observation? Curious how they compare with other studies, like Orzech.