A Novel Consumer Purchase Behavior Recognition Method Using Ensemble Learning Algorithm

Table 13

Comparison of experimental results after adding noise.

Noise ratio (%)	Algorithm	Index
		Accuracy	Precision	F1
1	LR	0.7436	0.7014	0.7334
	SVM	0.7603	0.7113	0.7485
	FSVM	0.7820	0.7548	0.7664
	RF	0.7716	0.7284	0.7548
	XGBoost	0.7837	0.7579	0.7767
	AdaBoost-SVM	0.7917	0.7553	0.7798
	AdaBoost-FSVM	0.8398	0.7910	0.8012
3	LR	0.7408	0.6822	0.7171
	SVM	0.7510	0.6904	0.7265
	FSVM	0.7796	0.7439	0.7564
	RF	0.7549	0.7045	0.7352
	XGBoost	0.7754	0.7370	0.7565
	AdaBoost-SVM	0.7801	0.7411	0.7688
	AdaBoost-FSVM	0.8292	0.7839	0.7925
5	LR	0.7207	0.6688	0.7010
	SVM	0.7321	0.6719	0.7063
	FSVM	0.7559	0.7359	0.7446
	RF	0.7338	0.6987	0.7123
	XGBoost	0.7530	0.7121	0.7321
	AdaBoost-SVM	0.7645	0.7207	0.7448
	AdaBoost-FSVM	0.8179	0.7764	0.7824