A Mental Workload Evaluation Model Based on Improved Multibranch LSTM Network with Attention Mechanism

Table 6

4-classification experimental results.

Model\Index	Acc	P	R	F-score
CNN	0.7553 ± 0.0946	0.7634 ± 0.1120	0.7717 ± 0.1162	0.7532 ± 0.1006
RNN	0.7721 ± 0.1102	0.7842 ± 0.1089	0.7586 ± 0.1412	0.7446 ± 0.1144
LSTM	0.7904 ± 0.1085	0.8008 ± 0.1092	0.7783 ± 0.1154	0.7678 ± 0.1019
Reference [31]	0.8242 ± 0.0890	0.8353 ± 0.1145	0.8126 ± 0.1093	0.8086 ± 0.0974
Reference [32]	0.8199 ± 0.0686	0.8321 ± 0.0854	0.8069 ± 0.0765	0.8027 ± 0.0679
Proposed	0.9264 ± 0.0687	0.9347 ± 0.0535	0.9110 ± 0.0846	0.9004 ± 0.0456