Introduction

In the previous chapter, you saw how to build a binary classifier using the famous Logistic Regression algorithm. A binary classifier can only take two different values for its response variables, such as 0 and 1 or yes and no. A multiclass classification task is just an extension. Its response variable can have more than two different values.

In the data science industry, quite often you will face multiclass classification problems. For example, if you were working for Netflix or any other streaming platform, you would have to build a model that could predict the user rating for a movie based on key attributes such as genre, duration, or cast. A potential list of rating values may be: Hate it, Dislike it, Neutral, Like it, Love it. The objective of the model would be to predict the right rating from those five possible values.

Multiclass classification doesn't always mean the response variable will be text. In some datasets, the target variable may be encoded into a numerical form. Taking the same example as discussed, the rating may be coded from 1 to 5: 1 for Hate it, 2 for Dislike it, 3 for Neutral, and so on. So, it is important to understand the meaning of this response variable first before jumping to the conclusion that this is a regression problem.

In the next section, we will be looking at training our first Random Forest classifier.