In this project, we will build a text message spam detector using a dataset of messages already classified by humans as either spam or not. We will use two methods to classify new messages as spam or not: Naive Bayes and Logistic Regression.
The Naive Bayes algorithm, which will will build from scratch, will analyze text messages by determining whether the words in the message appeared more often in spam messages or non-spam messages. It is called Naive Bayes because the algorithm does not try to interpret how any word in a message is being used - it only looks for what words are present in the message.
The Logistic Regression model will also use training data that consists of messages that are already labeled as spam or not to create a Sigmoid function that will determine if any given message is likely to be spam or not. Logistic Regression is often used when a dataset needs to be classified into two or more categories (such as spam or not spam).
Some of the messages in our dataset contain adult content.