Feature scaling

Feature scaling is a fancy term that means making all the numbers in a set more similar in size so that they can be compared more fairly with one another. Imagine you and your friends are trying to see who is the tallest. If your friend is 6 feet tall and you are only 3 feet tall, it wouldn't be fair to compare your heights directly because your friend is much taller.

The same goes for data. Let's say we're looking at a set of data about houses, such as their size in square feet, number of bedrooms, and price. In this set, the size of the house is measured in the thousands of square feet, the number of bedrooms is a small number, and the price could be hundreds of thousands or even millions of dollars.

If we wanted to make a comparison between the size of the house and the number of bedrooms, we would have to scale the features so that the numbers were more similar in size. We could do this by dividing the size of the house by 1000, and multiplying the number of bedrooms by 10. This would make the numbers more comparable and fair to compare against one another.

Feature scaling helps us make sense of data by making it more understandable and easy to compare. It helps us to avoid unfair comparisons that might not accurately reflect the true nature of the data.