Data-Driven Security: Analysis, Visualization and Dashboards

Chapter 9

Demystifying Machine Learning

“They know enough who know how to learn.”

Henry Adams

There are two types of people in information security—those who are completely intimidated by machine learning and those who know machine learning largely solved the spam problem and are completely intimidated by machine learning. It's easy to be intimidated when machine learning is described as “a type of artificial intelligence that provides computers with the ability to learn without being explicitly programmed” by TechTarget. (http://whatis.techtarget.com/definition/machine-learning). How can a computer do anything without being explicitly programmed? Or better yet, consider this rather well known definition from Tom M. Mitchell in his 1997 book titled Machine Learning:

A computer program is said to learn from experience E with respect to some class of tasks T and performance measure P, if its performance at tasks in T, as measured by P, improves with experience E.

Are you clear now on what machine learning is? This broad definition doesn't help much because it only describes the abstract results of machine learning, not what it is or how to use it. To help you understand machine learning at a practical and concrete level, we start this chapter with a learning task associated with realistic data. Prepare for the examples in this chapter by setting the directory to the working directory for this chapter and make sure the R libraries are installed (Listing 9-0).

Listing 9-0

# set working ...