What Is a Method That Can Be Used to Minimize

Multiple Choice

What is a method that can be used to minimize the penalized least squares criterion when using neural network modeling?

A) Fit the model 10-50 times using differing sets of randomly selected starter values to each fit.
B) Fit the model once with a predetermined starter value.
C) Use the set of parameters that yields the largest value of the penalized least squares.
D) Fit the model hundreds of times using differing sets of randomly selected starter values to each fit, to mimic using the entire population of data.