A Deep Learning-Based Framework for Human Activity Recognition in Smart Homes

Table 5

Results of all models in the “feature selection approach.”

Feature selection approach (50 features)	Accuracy (%)	F-measure (%)	Training time (sec)	Testing time (sec)
Baseline-ANOVA-RF	95.10	95.30	4.94	0.83
ANOVA-MLP	95.37	95.39	46.3	3.08
ANOVA-RNN	95.21	95.27	106.71	2.96
ANOVA-LSTM	95.27	95.30	135.04	3.62
ANOVA-GRU	95.14	95.20	142.87	3.67