Table 3.

Accuracy of 8 algorithms.

	Calculation method	Calculation formula	Accuracy (%)
Main symptom weight (for the first 3 symptoms, if the disease has cooccurrence, multiply it by the weight coefficient)	Add prior probability	P(DiseaseX|SymptomA,  SymptomB,  SymptomC)=P(SymptomA|DiseaseX)/P(SymptomA)∗P(SymptomB|DiseaseX)/P(SymptomB)∗P(SymptomC||Disease X)/P(SymptomC)∗P(DiseaseX) (for the first 3 symptoms, if the disease has cooccurrence, multiply it by the weight coefficient)	73
	No prior probability	P(DiseaseX|SymptomA,  SymptomB SymptomC)/P(DiseaseX)=P(SymptomA|DiseaseX)/P(DiseaseA)∗P(SymptomB|DiseaseX)/P(SymptomB)∗P(SymptomC|DiseaseX)/P(SymptomC)	69
	Log function form	log(PDiseasex1|SymptomA,  SymptomB,  SymptomC)=logP(SymptomA|DiseaseX1)/P(SymptomA|DiseaseX2)+logP(SymptomB|DiseaseX1)/P(SymptomB|DiseaseX2)+logP(SymptomC|DiseaseX1)/P(SymptomC|DiseaseX2)+logPDiseaseX1/PDiseaseX2	77
	Equal probability	P(DiseaseX|SymptomA,  SymptomB,  SymptomC=P(SymptomA|DiseaseX)/P(SymptomA))∗P(SymptomB|DiseaseX)/P(SymptomB)∗P(SymptomC|DiseaseX)/P(SymptomC)∗P(DiseaseX)	84.2
Main symptoms are not weighted	Add prior probability	P(DiseaseX|SymptomA,  SymptomB,  SymptomC)=P(SymptomA|DiseaseX)/P(SymptomA)∗P(SymptomB|DiseaseX)/P(SymptomB)∗P(SymptomC|DiseaseX)/P(SymptomC)∗P(DiseaseX)	73
	No prior probability	P(DiseaseX|SymptomA,  SymptomB,  SymptomC)/P(DiseaseX)=P(SymptomA|DiseaseX)/P(SymptomA)∗P(SymptomB|DiseaseX)/P(SymptomB)∗P(SymptomC|DiseaseX)/P(SymptomC)	67
	Log function form	log(PDiseasex1|SymptomA,  SymptomB,  SymptomC)=logP(SymptomA|DiseaseX1)/P(SymptomA|DiseaseX2)+logP(SymptomB|DiseaseX1)/P(SymptomB|DiseaseX2)+logP(SymptomC|DiseaseX1)/P(SymptomC|DiseaseX2)+logPDiseaseX1/PDiseaseX2	76
	Equal probability	P(DiseaseX|SymptomA,  SymptomB,  SymptomC)=P(SymptomA|DiseaseX)/P(SymptomA)∗P(SymptomB|DiseaseX)/P(SymptomB)∗P(SymptomC |DiseaseX)/P(SymptomC )∗P(DiseaseX)	83