Product developer’s guide to model training: Scaling numerical data

TLDR

While machines may have an easy time understanding numbers, it means nothing if there’s no meaning behind them. Learn techniques to scale to your numerical data, such as standardization and normalization to better grasp the correlation of your data.

Outline

• Recap
• Before we begin
• Numerical data types
• Why we scale
• Types of scaling
• Scaling with Pandas
• Use of numerical data

Recap

Last time, we looked at qualitative data, where we labeled the categories and assigned weights to data to make it machine readable. This time, we’re going to look at quantifiable data, aka numbers, and assign meanings to them through applying scales.

We’ll be looking at the different types of numerical data, discrete and continuous, as well as the algorithm or mathematical formula behind it. Finally, we’ll wrap it up by scaling all numerical values in our dataset.

Read more...