[Retracted] A Visual Human-Computer Interaction System Based on Hybrid Visual Model

Table 4

Algorithms with relatively high accuracy are extracted for further evaluation, including accuracy, F1-score, and recall. SVM (RBF) has the best performance among various algorithms.

Calssification method	Manhattan k-NN	Entropy decisio tree	Gini random forest	Gaussian kernel SVM (RBF)	Sigmoid NN	ReLu NN
Recall	0.711495	0.694532	0.695101	0.740983	0.671318	0.660655
Precision	0.716116	0.677113	0.745203	0.793791	0.766872	0.764888
F1-score	0.694213	0.664086	0.704173	0.744384	0.700354	0.691554
Accuracy	0.844735	0.824281	0.852715	0.875404	0.858406	0.852071