I. Introduction

Dimensional analysis is a technique used in physics to ensure an equation is consistent. For example, suppose we have an equation given by

Where _x _is the distance (feet or meter) and t is time (second). For this equation to be consistent, A must have the units of distance (meter), while B must have the unit of distance/time (meter/second).

In machine learning, we often deal with datasets containing features with different dimensions. For example, a cars dataset can have features such as length of a car (feet or meters), mass of a car (pounds or kilograms), **age of a car **(years), fuel consumption rate (miles per gallon or kilometers per liter), the color of a car (green, red, blue, black, white, etc.). Since features come in different units, it is important to be careful when performing data analysis as we can’t add apples and oranges. For example, you can’t add your **mass of car **feature column to your fuel consumption rate column, this would be a meaningless operation.

In this article, we discuss the importance of dimensional analysis in machine learning.

#machine-learning #standardization #feature-scaling #data-science #linear-regression