Application of sample balance-based multi-perspective feature ensemble learning for prediction of user purchasing behaviors on mobile wireless network platforms

EURASIP Journal on Wireless Communications and Networking

Table 2 P, R, and F1 scores of eleven user purchasing behavioral prediction models

Evaluation index		P	R	F1 score
Model type		P	R	F1 score
Comparison model	co-EM-LR	0.0734 ± 0.0025	0.0697 ± 0.0026	0.0715 ± 0.0046
	GDBT	0.0822 ± 0.0010	0.0789 ± 0.0035	0.0808 ± 0.0044
	MMSE	0.1221 ± 0.0041	0.0983 ± 0.0064	0.1089 ± 0.0055
	LR-XGBoost	0.1104 ± 0.0032	0.1048 ± 0.0024	0.1075 ± 0.0045
	RNN	0.0874 ± 0.0047	0.0783 ± 0.0054	0.0826 ± 0.0041
	MLP-LSTM	0.0852 ± 0.0019	0.0708 ± 0.0065	0.0773 ± 0.0021
Ablation model	FCVS without XL	0.1704 ± 0.0054	0.1311 ± 0.0012	0.1482 ± 0.0042
	FCVS without LL	0.1652 ± 0.0036	0.1458 ± 0.0024	0.1549 ± 0.0012
	FCVS without CDC	0.1557 ± 0.0054	0.1326 ± 0.0041	0.1432 ± 0.0028
	Stacking	0.1502 ± 0.0029	0.1835 ± 0.0050	0.1652 ± 0.0036
	FCVS	0.1714 ± 0.0013	0.1871 ± 0.0027	0.1789 ± 0.0024