![]() ![]() To find this minimum we can set the derivative equal to zero and solve for m.īut what does it mean to take a derivative when it comes to vectors? If I assume x and y are scalars and do not specify the derivative of the transpose, then I can use the product rule:Īnd then I can use the product rule (df/dm * g + f * dg/dm), but the result looks nothing like the answer which is supposed to be m=y Tx/x Tx. Minimizing (y-mx) T(y-mx) over m corresponds to finding the estimate of m that minimizes the squared error cost function. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |