Line vector, shape( 1, n), n is the dimension of the vector; The row vector can be simply written as a vector with 1 rows and n columns.

A column vector with shape = (n, 1), where n is the dimension of the vector; The column vector can be simply written as a vector with the number of columns of 1 and the number of rows of n;

The dot product of the row vector and the column vector is as follows:

Therefore, in order to simplify the implementation of the program, "transforming the original data into a matrix" can realize the neuron calculation of m samples and n eigenvalues (weighted sum+activation function).