Understanding This NN Training Example

Improve Neural Network Generalization and Avoid Overfitting

This is what I am attempting to work through. What I of course want, is x amount of Neural Networks trained, then getting the average output for new inputs.

I get two items from following the Multiple Neural Networks tutorial that I am unsure about. What does "neti"(network) and "nets"(array) mean? How do I then use this on new outputs?