Knowledge-Based Recurrent Neural Network for TCM Cerebral Palsy Diagnosis

Table 1

The classification accuracy of the KBRNN and baselines.

	Preset	Preset + 50% train-set	Preset + 100% train-set
KBRNN	79.31	82.03	83.12
RNN	44.28	71.32	78.03
LSTM	45.01	76.04	79.49
GRU	45.19	76.22	79.85
Bi-RNN	44.64	76.59	82.39
Bi-LSTM	45.74	77.13	80.40
Bi-GRU	45.91	79.85	82.58
CNN	47.37	79.67	81.67
DAN	48.09	80.21	81.85