Learning Users’ Intention of Legal Consultation through Pattern-Oriented Tensor Decomposition with Bi-LSTM

Table 2

F1 score of algorithms based on multiple deep neural networks.

Optimization	Basic algorithm
	TextCNN	TextCNN attention	TextRNN	TextRNN attention	LSTM	BiLSTM	GRU	Bi-GRU
The original method	0.64	0.70	0.68	0.71	0.70	0.74	0.71	0.77
CP TP method	0.72	0.77	0.80	0.78	0.83	0.87	0.83	0.84
Tucker TP method	0.75	0.77	0.76	0.81	0.80	0.84	0.82	0.86
Pattern-oriented TP	0.83	0.85	0.84	0.86	0.88	0.92	0.86	0.89