As I understand a vector is anything with a dimension of n x 1. While a matrix is anything with a dimension of n x n. Where n can be any number.

So, to my knowledge, when we convert a scalar repetitive operation into a single matrix, this is called vectorization. I'll take the example of linear regression, hypothesis function. We convert H(x)=theta0+theta1x into two matrices, X, with a dimension of m x 2, and Theta with a dimension of 2 x 1, where m is the number of training examples. Here one vector is formed, while the other is a matrix. So wouldn't a better term be "matrixiation".

Why have we named it vectorization?

  • 1
    $\begingroup$ Matrices can have dimension $n \times p$ -- that is, matrices don't have to be square. $\endgroup$
    – Sycorax
    Commented May 10, 2020 at 23:28
  • 4
    $\begingroup$ All matrices are a fortiori vectors. The term "vectorization," though, comes from computer science, not stats or mathematics. $\endgroup$
    – whuber
    Commented May 11, 2020 at 12:17


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.