Personalized Body Constitution Inquiry Based on Machine Learning

Table 2

Experimental performances of different models.

Body constitution combination	Classifier	k	Acc (5-folds) (%)	Micro-averaging			Macro-averaging			Response time (s)
				Precision (%)	Recall (%)	F Score (%)	Precision (%)	Recall (%)	F Score (%)
Balanced Constitution; Yang-deficient Constitution; Qi-deficient constitution	LDA	16	99.88	99.88	99.88	99.88	99.86	99.88	99.87	0.009
	ANN	18	99.90	99.90	99.90	99.90	99.89	99.90	99.90	0.524
	SVM	15	99.76	99.76	99.76	99.76	99.74	99.75	99.75	0.013
	RF	15	87.29	87.29	87.29	87.29	87.81	86.34	86.38	0.036
	KNN	14	97.18	97.18	97.18	97.18	97.08	97.05	97.06	134.273
Yang-deficient Constitution; Phlegm-dampness Constitution; Stagnant blood constitution	LDA	22	99.84	99.84	99.84	99.84	99.83	99.84	99.84	0.006
	ANN	25	99.81	99.81	99.81	99.81	99.81	99.82	99.81	0.389
	SVM	28	96.02	96.02	96.02	96.02	96.28	96.01	96.03	0.011
	RF	18	66.63	66.63	66.63	66.63	67.49	65.95	65.66	0.049
	KNN	22	87.48	87.48	87.48	87.48	87.45	87.50	87.46	1710.521