 However, when dealing with such a large number of values and operations, computers tend to perform more efficiently with matrices. I'll use the convention that any matrix or vector is bold-faced in order to distinguish them from scalar quantities as shown. From our multiple regression equation, we can see that y is equal to x-beta plus epsilon, and y-hat is equal to x-beta hat. Here, y is an n cross 1 matrix, or an n-dimensional column vector. x is an n cross p plus 1 matrix. It's p plus 1 and not p because of the additional intercept multiplicand which we take as 1. Beta is a p plus 1 dimensional column vector, and epsilon is an n-dimensional column vector.