On the planet of knowledge and computer programs, the concept of Machine Learning might sound like a tricky nut to crack, filled with tricky math and sophisticated ideas.

This is the reason today I would like to decelerate and take a look at the fundamental stuff that makes all this work with a brand new issue of my MLBasics series.

Today’s agenda is giving our good old Logistic Regression a swanky upgrade.

Why?

By default, Logistic Regression is restricted to two-class classification problems. Nevertheless, we regularly face multiple-class problems.

So let’s dive into the fascinating world of leveling up Logistic Regression to have the opportunity to sort things into greater than two baskets 👇🏻

Within the ML field, Logistic Regression stands as an optimal model for binary classification problems.

It’s the trusted path towards decision-making.

Nevertheless, there’s an enormous problem with Logistic Regression: It is sort of a coin toss — heads or tails, A or B.

But what if you will have multiple classes?

Logistic regression will not be enough to handle a multiple-class classification. Subsequently, to perform so, the model must be adapted and there are two principal options:

• The primary easy approach is using multiple Easy Logistic Regression models to discover each considered one of the classes we would like. It is a simple solution.
• A second approach is to generate a brand new model that accepts multiple classes.

So let’s break down each approaches: