Member-only story
Classification Metrics Data People Should Know
Learn the foundations for data science success!
Learning classification metrics can be a foggy mess. ☁️ There are lots of metrics. Many sound alike. Sometimes there are two names for the same thing. Sometimes there are five. 🤦♀️🤦♂️
This series of articles is designed to lift the fog.🌤 It’s a guide to help you understand, use, and remember the seven most common classification metrics.
Today we’ll explore the foundation that everyone needs. Data scientists and statisticians must have the foundation down cold before moving on to more advanced metrics. 😀
Let’s get to it!🚀
In binary classification problems there are two possible outcomes: true or false, 1 or 0, buy or don’t buy, infected or not infected, etc. In this article we’ll focus on binary classification, because it’s common and this stuff is confusing enough with just two outcome classes. 😉
The metrics we’ll look at are in the scikit-learn Python package. Each function is imported from the sklearn.metrics module.
We’ll use a few hypothetical classification tasks to illustrate how these metrics work. Let’s start with predicting whether a website visitor would purchase a shirt at Jeff’s Awesome Hawaiian Shirt store…
