Couldnt find a AI live chat, this is kinda out of topic but does anyone of you know how logistic regression works ?

I mean, I know how it works to the level I can glean from typing "logistic regression tutorial" into a search engine, but I assume you've already done that. Is there a specific snag you've hit that we can help you with?

@DMGregory Yes ^^ Well we also had a lecture, but my prof somehow does not reply to emails since days :D Our task was to implement it from scratch, mostly by using matrixes. The formula for optimizing the weights is looking like this in our lecture : prnt.sc/226zp12

I actually implemented this 1:1 using c++ and eigen. The problem however is that there seems to be an error in either my code or the formula itself. I also couldnt find any similar formula online

( w = weight , alpha = learn-rate ) forgot to mention this

pastebin.com/QVt3aBX1 The code... basically the formula shown above. However this produces an weird output, the gradient doesnt really change at all and the weights grow and grow and grow. The costs however increase which doesnt make sense either.

The prediction implementation however looks correct. And the way i implemented the weight optimization formula looks also correct... any idea how i may figure out where the mistake is ?

@Tyyppi_77 Thanks im gonna look at that ^^ I think those few lines above are actually the formula for the gradient descend, which is used to update the weights. But i assume there multiple ways