A classifier assigns inputs into one of the N classes based on some properties (also known as features) of inputs. Classifiers have widespread applications, such as e-mail spam filtering, finding the most promising products, selecting customers for closer interactions, and taking decisions in machine learning situations. Let's explore how to implement a classifier using a large dataset. For instance, a spam filter will assign each e-mail to one of the two clusters: spam mail or not spam mail.

There are many classification algorithms. One of the simplest, but effective, algorithm is the naïve Bayesian classifier that uses the Bayes theorem involving conditional probability.

In this recipe, we will also focus on the Amazon metadata dataset as before. We will look at several features of a product, such as the number of reviews received, positive ratings, and known similar items to identify a product with potential to be within the first 10,000 sales...