Are neural networks best used to help logistic regression? An example from breast cancer survival analysis

摘要: Artificial neural networks are popularly used as universal nonlinear inference models. However, they suffer from two major drawbacks. Their operation is opaque because of the distributed nature representations form, and this makes it different to interpret what do. Worse still, there no clearly accepted models generality which difficult demonstrate reliability when applied future data. In paper generate hypotheses concerning interaction terms integrated into standard statistical that linear in parameters, where significance model, can be assured using well established tests.