How the matrix (W) is [4*3] works?


strong text

Because in that layer, it has 4 neurons, and each neuron accepts 3 features as its input.

Raymond

1 Like


Actually I was asking how this matrix is (4x3), because if it was multiplied by input features then it can be (4x3) like
[[w1[1]Tx1 w1[1]Tx2 w1[1]Tx3]
[w2[1]Tx1 w2[1]Tx2 w2[1]Tx3]
[w3[1]Tx1 w3[1]Tx2 w3[1]Tx3]
[w4[1]Tx1 w4[1]Tx2 w4[1]Tx3]]
then it can be (4x3) matrix but here before multiplication how it can be (4x3) matrix.

Hello Priyanshu,

Welcome to the community.

The 4x3 that you are seeing is just an explanation given by Prof Ng to represent the value z[l] here.
4x3

In addition to the great explanations and diagrams from Raymond and Rashmi, here’s another thread from a while back that goes through this more at the level of math formulas.

@priyanshu_Kumariitp

What else should it be if not a 4×3 matrix?