I have finished week2 assignment and in that I have calculated cost function as -1/m * (np.sum(Y * np.log(A) + (1-Y) * np.log(1-A))) where we do a normal multiplication of Y and log(A) but in week3 assignment while calculating cost function why is it that we use np.multiply followed by np.sum or np.dot instead of normal multiplication?
The input and output in both assignments are array then what makes the difference here?
np.multiply is the same as “*”. They are two ways to say the same thing. By analogy np.dot can be abbreviated as “@”, although Prof Ng chooses not to use that notation.
For the cost computation, the fundamental operation is multiplying the corresponding elements of two vectors and then adding up the products to get a scalar. You can do that in two steps with np.multiply followed by np.sum or you can do it in one step using np.dot. Well, I guess you could say np.dot requires two steps also, because you need a transpose on one of the vectors first.
The cost value is a scalar, but some of the methods will produce a 1 x 1 array. The grader seems not to care about this difference.