Okay, so let's look at logistic regression curves and think about how r is fitting the data. So here I'm going to use the package manipulate. And i have my beta2 and my beta1, set here. And so here's my logistic regression curve, looks like an s kind of, and here's beta1. So let's see what happens as I render it. Or as I change it. Notice as I get it closer up to higher values, it becomes more peaked, okay? And then if I go negative, it actually flips around, okay? And then my beta0 curve, what happens to this? It just shifts it along. It's hard to tell because I'm keeping the don't. The x-axis here is always the same, but it's just shifting the s-curve to the left or to the right. Let's think about what r is doing when it's trying to fit logistic regression curves. Imagine I have a bunch of zeros and ones, on my x-axis one is up here, and zero is down there. So imagine for this regressor, this x-regressor here, I have a bunch of zeros down there, but maybe and an occasional one, some, but then over here, I have a lot more ones, and then right here I have some zeros, okay? So what r is trying to do, okay? So it knows as I head, we can see as I head to larger x values, it become increasingly likely that my outcome is a one. As I head towards smaller x values, it becomes increasingly likely that my outcome is a zero. So what r is trying to do is trying to move the s-curve. It's trying all different sorts of s-curves, right? All different sorts of logistic curves to find the one that best matches up with the associated probabilities, okay? And that's all that logistic regression is trying to do. And it does it with the principle of so-called maximum likelihood. Which again is something if you take the inference class from the data science specialization, you'll know a little bit more about. So just imagine these points staying there, and I'm going to get rid of them, okay? And then just see what r has under its disposal is to be able to move these two sliders, and then define the s-curve that fits that data the best, okay? And so that's all that logistic regression is doing. And this function, this the fitted function out, right? This is the function. E to the Beta not, plus beta 1x over one, plus e to the beta not, plus beta 1x, that's the logistic regression model but converted back to the probability scale, okay? So, I hope this helps you understand a little bit of behind the scenes, what logistic regression is trying to do.