Analyzing Customer Conversions with Machine Learning and Decision Trees

Here we will expand your knowledge of explanatory analysis and show you how to use decision trees to understand the drivers of consumer behavior. We will start by comparing and explaining the differences between logistic regression and decision tree models, and then discuss how decision trees are built and trained.

Logistic Regression Vs Decision Trees
A logistic regression model learns from the data by finding the linear combination of feature variables that best estimates the logistic probability of an event occurring. Decision trees, as their name suggests, learn from data by growing a tree.

We'll discuss how decision tree models grow and how to build trees in more detail below, but first... The main difference between logistic regression and decision tree models is the fact that logistic regression algorithms look for a single best linear boundary in the feature set, while the decision tree algorithm splits the data to find the subgroups of data that have a high probability of an event occurring. It'll be easier to explain this with an example. Let's take a look at the following diagram:

https://www.narrativetext.co/the-linear-model/analyzing-customer-conversions-with-machine-learning-and-decision-trees