# Linear Algebra for Machine Learning | by Luay Matalka | Dec, 2020

0
48 When coping with actual numbers, 1 is the identification for multiplication. In different phrases, for any actual quantity z, the #1 occasions z will equivalent z occasions 1, which can equivalent z. So the identification assets of one implies that any actual quantity z multiplied by 1 is the same as z, thus permitting z to stay its identification.

We even have identification matrices. In different phrases, for any matrix A, there can be an identification matrix, I, that once multiplied with matrix A, equals to matrix A.

AI = IA = A

Identity matrices are sq. matrices, which means that the selection of rows equals the selection of columns. Identity matrices are denoted by I, infrequently as Inxn, with nxn being the measurement of the identification matrix. For a matrix of the measurement mxn, its identification matrix would be the measurement nxn, which must make sense, since for the ones matrices to be multiplied, the selection of columns of the primary matrix will have to equivalent the selection of rows in the second one matrix.

Examples of identification matrices:

As we will see above, identification matrices (with the exception of for when n=1), have 1’s alongside the diagonals, and zeros far and wide else.

As mentioned ahead of, generally, matrix multiplication isn’t commutative, until this is a matrix multiplied by its identification matrix.