Distant Supervision with Transductive Learning for Adverse Drug Reaction Identification from Electronic Medical Records

Table 3

The list of parameters for assessment.

Parameter group	Parameter type	Parameter subtype	Parameter name	Variable name
Document representation		Syntactic lemmatization	Syntactically lemmatized lexicon	L
			Surface lexicon	S
		Pattern granularity	Phrase form	P
			Word form	W
	Pattern-weighting models	Bernoulli	Binary	B
		Multinomial	TF (term frequency)	TF
			TFIDF (TF-inverse document frequency)	TFIDF
Model assumption	Independent assumption	EM with naïve Bayes		iEM
	Dependency representation assumption	EM with Bayesian network		dEM
Model decision method	Soft decision making			S
	Hard decision making			H
Learning method	Supervised learning			SL
	Transductive learning	Initial weight method for unlabeled data	Supervised model
			Equal probability
			Random probability