Machine Learning Introduction

Let’s start with a little terminology. According to Wikipedia, machine learning (ML) is a class of artificial intelligence methods characterized by their not providing direct solutions to problems but rather training systems to apply solutions.

There are many methods of machine learning, but they can roughly be divided into two groups: learning with a teacher and learning without a teacher.

In the case of learning with a teacher, a person supplies the machine with initial data in the form of situation–solution pairs. The machine learning system then analyzes these pairs and learns to classify situations based on known solutions. For example, a system can learn when to mark incoming messages as spam.

In the case of learning without a teacher, the machine receives unsorted information — situations — without solutions and learns to classify those situations based on similar or different signs without human guidance.
Machine learning in online marketing
Marketers use machine learning to find patterns in user activities on a website. This helps them predict the further behavior of users and quickly optimize advertising offers.
What is the potential of behavioral data?
In psychology, a pattern is a particular set of behavioral reactions or a common sequence of actions. Therefore, we can talk about patterns with regard to any area where people use templates (which is most areas of life).

Consider the example of a pattern used on websites. If the user isn’t interested in the offer in the pop-up window shown below, they can close this window by:

clicking on the X sign
clicking No thanks
clicking anywhere on the site that’s outside the pop-up window.