liveBook · Manning

Spam

As explained in Grokking Machine Learning

An unwanted email, including unsolicited promotions or chain letters.

Spam is junk or unwanted email, such as unsolicited promotions and chain letters. Non-spam email is often called ham. section 1.3

Spam detection classifies email as spam or ham. A naive Bayes model estimates these probabilities from email features; one example uses words, where “lottery” is more likely in spam than “meeting.” section 8.2

Estimating spam probability

In a sample of 100 emails, 20 are spam, giving a prior estimate of 0.2. Among 20 emails containing “lottery,” 15 are spam, so the estimated probability of spam given that word is 0.75. section 8.2

How Grokking Machine Learning covers it

How deep it goes
Develops a word-based probability example, including prior and conditional estimates, without detailing a complete filtering implementation.
Examples
Uses chain letters, promotions, a sender’s email history, and the words “lottery” and “meeting.”
What you can do after
Apply probability estimates to classify email as spam or ham and direct likely spam to a spam folder.
How it connects
Connects spam with ham, machine learning, naive Bayes, and probability.
Cover of Grokking Machine Learning
liveBookmanning.com
Read Grokking Machine Learning on liveBook Get the book at manning.com

Also explained in

Compare how 3 books explain spam →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Grokking Machine Learning. The section links point to the parts of the book it draws on.