Parameter Optimization and State Evaluation of Basketball Teaching Based on BPNN

Table 2

Comparison of errors of different training functions.

Hidden layer function	Output layer function	Error percentage (%)	Mean square error
logsi	logsi	351.62	181.34
logsi	tansig	40.62	0.92
logsi	purelin	0.07	0.001
tansig	logsi	340.37	162.34
tansig	tansig	31.71	1.34
tansig	purelin	1.61	0.01
purelin	logsi	341.28	145.28
purelin	tansig	119.28	112.94
purelin	purelin	197.29	99.35