Is there a classical regression model where, for 𝑖=1,…,𝑛,

𝐸(𝑌ᵢ) = 𝑁 𝑝ᵢ

with 𝑁 a known constant, and

𝑝ᵢ=(exp 𝑋ᵢ β) / (∑ⱼ exp 𝑋ⱼ β)

Thus 𝑝ᵢ ∈ (0,1) and ∑ 𝑝ᵢ = 1.

Note that this is *not* a multinomial logistic regression. There is a single vector β to estimate. It should be estimated from a single set of observations 𝑌₁,…,𝑌ₙ (and the covariates 𝑋₁,…, 𝑋ₙ).

#statistics #statisticalmodel #glm #rstats

This is an example that compares fitting a #logistic regression as a #generalized #linear #model #GLM & as an #AI #MachineLearning using #tensorflow. Marvel at the computational waste (number of epochs ; a typical IWLS #GLM fit converges in <10 iterations) https://atm.amegroups.org/article/view/30334/html

GLB qui s'en prend publiquement à une chercheuse universitaire dont le seul "crime" est de porter le voile en faisant le parallèle avec le HAMAS...que dire ?
#GLM #Belgique #Voile

imnotycobb: Since Jan 1 there have been 30 (now 31) bolides or meteors captured with #GOES #GLM lightning mapper. Note: this data only goes through the 8th since each event is verified for accuracy. More info about bolide detection can be found here #okwx https://twitter.com/imnotycobb/status/1616468906100621314

Thread from @AndyChenML@twitter.com arguing that #GLM-130B beats GPT-175B, is open source, and can run on a single A100 with 4-bit quantization: #llm https://twitter.com/andychenml/status/1611529311390949376?s=46&t=jgk40-JPcPZRoMLFx1B0tw

Cheatsheet for linear regression in r. #rstats #linearRegression #glm #cheatsheets
I’m trying not top post things from the bad place but this one is really good!
Originally posted by Ben Larson blarson424 on the bird site.