can we use neural networks like regression models for predicting a parameter while giving multiple other parameters as features ?
if yes, what type of activation function and cost function should I use?
thank you
Yes you can use to predict one parameter if that parameter has a dependency on the other parameters, activation and cost depends on the application and that dependency. In tensor flow most of these steps are straightforward to set, you should take some specializations here to get some info about these. You can start with the Machine Learning Specialization as generic knowledge about ML.
1 Like