Medical Text Classification Using Hybrid Deep Learning Models with Multihead Attention

Table 15

Performance comparison of accuracy (%) of the proposed deep learning models with other deep learning models on the same dataset.

Methods	Hall mark dataset	AIM dataset
CNN	68.55	82.17
LSTM	70.76	83.16
BiLSTM	72.58	87.77
CNN-LSTM	71.81	91.98
CNN-BiLSTM	73.99	93.06
Logistic regression	61.91	72.92
NBC	65.35	73.84
SVM	66.99	84.55
BiGRU	69.34	89.98
Proposed method 1: quad channel hybrid LSTM model	75.98	96.72
Proposed method 2: hybrid BiGRU with multihead attention model	74.71	95.76