Classification of FinTech Patents by Machine Learning and Deep Learning Reveals Trends of FinTech Development in China

Table 5

The classification performance of the different models on the test set.

Category	Metric	SVM	KNN	NB	DT	CNN	RNN	LSTM	BERT
FinTech-unrelated	Pr	0.74	0.54	0.47	0.62	0.81	0.47	0.44	0.84
	Re	0.67	0.76	0.34	0.6	0.87	0.32	0.56	0.9
	F1	0.71	0.63	0.40	0.61	0.84	0.38	0.50	0.87
Insurance	Pr	0.58	0.45	0.46	0.34	0.82	0.42	0.37	0.81
	Re	0.30	0.56	0.22	0.32	0.90	0.30	0.34	0.95
	F1	0.39	0.50	0.30	0.33	0.86	0.35	0.35	0.87
Data analytics	Pr	0.57	0.63	0.67	0.63	0.78	0.67	0.65	0.74
	Re	0.73	0.76	0.54	0.71	0.58	0.54	0.59	0.55
	F1	0.65	0.69	0.59	0.67	0.67	0.59	0.62	0.63
Lending	Pr	0.92	0.62	0.85	0.80	0.84	0.62	0.87	0.86
	Re	0.68	0.72	0.44	0.74	0.98	0.42	0.66	0.96
	F1	0.78	0.67	0.58	0.77	0.91	0.50	0.75	0.91
Payment	Pr	0.76	0.62	0.40	0.59	0.90	0.62	0.66	0.90
	Re	0.74	0.78	0.24	0.70	0.86	0.42	0.61	0.78
	F1	0.75	0.69	0.30	0.64	0.88	0.50	0.63	0.83
Investment	Pr	0.90	0.84	0.78	0.82	0.82	0.88	0.76	0.77
	Re	0.72	0.64	0.64	0.74	0.82	0.60	0.74	0.82
	F1	0.80	0.73	0.70	0.78	0.82	0.71	0.75	0.8
Security	Pr	0.67	0.46	0.42	0.45	0.89	0.44	0.62	0.86
	Re	0.67	0.59	0.33	0.47	0.80	0.28	0.49	0.66
	F1	0.67	0.52	0.37	0.46	0.85	0.35	0.55	0.75
Average	Pr	0.72	0.64	0.55	0.63	0.84	0.59	0.65	0.83
	Re	0.72	0.61	0.54	0.63	0.83	0.57	0.64	0.80
	F1	0.71	0.61	0.52	0.63	0.83	0.55	0.64	0.81
	Acc	0.72	0.61	0.53	0.63	0.84	0.56	0.64	0.83

The significance of the symbol “” means that the model performs significantly better than others in the patent classification.