A mostly broken version of logistic regression that I was trying to program. It kind of not really works, but only sometimes. It might work if you tweak the learning rates of each of the weights and biases enough, because for some reason, the normalization didn't really work.
I was taking the Andrew Ng's course on machine learning and I wanted to recreate one of the lessons in scratch. I made the first part of this project in scratch 1.4 as a challenge, but then I realized that was way too hard, so I just switched to scratch 3.0 and fixed and finalized everything here.
