To know linear regression we need to understand why it came and what it does as a whole.
When we get any data set let’s say how many people are obese vs. inactive.
We will plot a graph where the x-axis is the percentage of people inactive whereas the y-axis is the percentage of people obese. What we will get is a graph full of points.
So linear regression is a machine learning algorithm that is used to predict future anomalies. So in order to predict we will draw a straight line based on current data and see if our line is closest to the data points that are already present. If it is then we can say that for the changing x-axis the point we will get in line will be the best possible prediction.
As it’s been seen the red line is predicted linear regression which is nearer to the data points and hence for the data points or let’s say future x value the predicted y value will be nearest to be accurate.
So in order to find this line we can use the squared error function(Cost Function) and minimize the value. The minimum value will be the line we are looking for.
Here hypothesis is the line equation we are looking for. Which looks like y=mx+c.