Given a set of data points (in multiple dimensions), logistic regression tries to fit a curve between the data points that best represent it. A more formal definition is; it is a technique that finds relationships between a set of independent variables and a dependent variable—here, the independent variables being the input data and the dependent variable being the labels corresponding to the data. The mathematics behind logistic regression involves probability. Given a data point, we calculate the probability of that data point belonging to a particular label.

Using logistic regression, let's try to build a digit classification program. Given an image of a digit, it will output what digit it is. For this task, we will ...