Understanding Simple Linear Regression

Intro to Regression Algorithms | Data series | Episode 4.1

This article is designed to teach the underlying theory of linear regression. How to code and implement this algorithm in Python will be included in episode 4.3. This article also covers some basic Data Science Terminology which is important to know for future episodes.

What is Simple Linear Regression?

Simple Linear regression is a common supervised machine learning algorithm (see episode 3) used by data scientists to make predictions for numerical data such as: future house prices or next year’s sales of a company.

Simple Linear regression should only be used on data where there seems to be a linear relationship between variables. That is our data follows a straight line.

Examples may include house size and house price, or customer reviews and company sales, or height and weight.

The reason why is that simple linear regression relies on a reasonable relationship between variables to make accurate predictions.

Overview

Take a look at the following graph looking at the Temperature and Humidity in Szeged, Hungary.