This is a mandatory model to know if you are performing a classification task using machine learning, because it is simple and is often used as a first benchmark to evaluate the performance of more complicated models. For binary classification, this model produces the conditional probability of the target belonging to the positive class. This model is another example of a parametric model; the learning algorithm will try to find the best combination (vector) of parameters 
such that the estimated probabilities produced by the equation are as follows:
We are close to 1 when the target belongs to the positive class and close to 0 when the target belongs to the negative class. So, by definition, this model predicts probabilities, and then we use this probability to predict classes. To understand it better, let's produce our first simple logistic regression...