Neural Network-Based Train Identification in Railway Switches and Crossings Using Accelerometer Data

Table 4

Number of layers and number of trainable parameters for the evaluated neural network architectures.

Model	MLP1	MLP2	MLP3a	MLP3b	CNN	LSTM
Number of layers	2	3	4	4	4	2
Number of trainable parameters	100 605	110 705	120 805	1 000 005	1 280 989	260 605