Apache Spark Deep Learning Cookbook

How it works...

This section explains how the cost function is used to generate weights and bias.

1. A for loop will be implemented that will perform gradient descent on the weights and bias to tweak the values until the cost function gets close to 0.
2. The loop will iterate over the cost function 100,000 times. Each time, a random value for height and weight from the 29 individuals is selected.
3.  A summation value, z, is calculated from the random height and weight, and the input is used to calculate a predictedGender score with the sigmoid function.
4. The cost function is calculated and added to a list that tracks all cost functions through the 100,000 iterations, allCosts.
5. A series of partial derivatives is calculated with respect to the summation ...