Analysis of Medical Opinions about the Nonrealization of Autopsies in a Mexican Hospital Using Association Rules and Bayesian Networks

Table 18

Presentation of the best results of each algorithm evaluation.

Association analysis
Data set	Algorithm	Parameters
C	Apriori	Confidence = 0.9
		Support = 0.4
		Rules = 15
	FPGrowth	Confidence = 0.9
		Support = 0.4
		Rules = 8
	PredictiveApriori	Support = 0.5
		Rules = 12
	Tertius	Support = 0.5
		Rules = 12
D	Apriori	Confidence = 0.9
		Support = 0.5
		Rules = 10
	Tertius	Support = 0.5
		Rules = 10
Classification analysis
Data set	Algorithm	Class	Search algorithm
D	BayesNet	reason_no_aut	Tan
		aut_reason	HillClimber
		fam_rejection	Tan
		years_pract	Tan
		area	Tan
		cases	Tan
		category	Tan
		com_sug_op	Tan
		spec_inst	HillClimber
		gral_med_inst	Tan
		method_request_aut	HillClimber
		arb_findings	Tan
		dem_findings	Tan
		disc_findings	Tan
		physician_reason_aut	Tan
		no_hosp	Tan
		staff_request_aut	Tan
		grade	HillClimber
Data set	Algorithm	Attribute selection measures
aut_reason	MultilayerPerceptron	InfoGain
reason_no_aut	MultilayerPerceptron	InfoGain
com_sug_op	MultilayerPerceptron	InfoGain