How to determine the number of neurons in a neural network?

Question

Neural networks

I want to write a neural network for detecting network anomalies. Feedforward neural network. The data represents 12 parameters inherent in incoming packets. Those. does this mean I should have 12 input neurons, 12 neurons in the hidden layer, and 2 neurons in the output?
In this case, each neuron in the input layer will receive one parameter from the packet?
Am I understanding everything correctly or should I do it differently?
I don’t know anything about this topic yet, and I haven’t found a specific practical answer yet either in the literature or on the forums. Tell me please.

Answer 1

It is better not to try to solve this problem as a classification, but immediately take up approaches from anomaly detection.

Answer 2

There is no formula for determining the size of hidden layers. Often it is determined by enumeration or based on common sense.