Posts

Showing posts with the label neural networks machine learning

Activation Functions In Neural Network

Image
It's  recommended  to  understand  what's a neural  network  before  reading  this  composition . In The  process  of  erecting  a neural  network , one of the  choices  you  get  to  make  is what activation  function  to  use  in the  retired   subcaste  as  well  as at the  affair   subcaste  of the  network . This  composition  discusses some of the  choices . Rudiments  of a Neural  Network - Input   Subcaste -This  subcaste  accepts  input  features. It provides  information  from the  outside   world  to the  network , no  calculation  is  performed  at this  subcaste ,  bumps   then   just   pass  on the  information  (features) to the  retired   subcaste . Hidden   Subcaste -  Bumps  of this  subcaste  aren't  exposed  to the  external   world , they're the  part  of the abstraction  handed  by any neural  network .  Retired   subcaste  performs all  kind  of  calculation  on the features  entered  through the  input   subcaste  and  transfer  the  result  to t