Linear Regression

Question 1

What is our ‘model’ in linear regression

Answer 1

Simply a line that goes through (in the middle of) our plotted data points

The ‘direction’ (I.e. angle/tilt) of this line is a representation of the data’s trend…
I.e. the features relationship with one another

E.g. positive correlation, negative, linear etc
And to what degree, e.g 35 degree positive correlation

Question 2

How is our model obtained?

Answer 2

Our model, aka the line, is derived via the: Ordinary least squares sum
2
(y - ŷ)^ >min

y = actual value of data (data point) 
ŷ = what the line predicts (where the line’s at on the x axis) 

Take the difference between y and ŷ, square it, and find the minimal possibility

So, all the distances from the plotted data points to the line are calculated.

y        y     
 :        :    y
———————- ŷ
   :       y
   y

We want a line that creates the least distance between it and all the plotted data points