Richard_Kennaway comments on A Simple Introduction to Neural Networks

[Richard_Kennaway]

One reason for using squared errors, which may be good or bad depending on the context, is that it’s usually easier to Do Mathematics on it.