Is there a standard systematic approach to determining the optimal Neural Network configuration for the specific data?

Question

Just wondering if there is a standard method or algorithm that can discern the optimal neural network configuration (i.e no of layers, size of layers and activation functions) f