Table 5.

Hypertension: Comparison of F1 measure for hypertension models between 2008, 2014, and combined 2008/2014 broken down by SVM kernel, weighting scheme, and feature space.

Kernel	Weight	Feature Space	2008	2014	Combined 2008/2014	% Change 2008, 2014	% Change 2008, Comb.	% Change 2014, Comb.
RBF	Count	Min, Lex & CUI	0.957	0.949	0.953	−0.84	−0.39	0.45
RBF	Count	Min & Lex	0.953	0.949	0.953	−0.45	0.00	0.45
RBF	tf-idf	Min & Lex	0.952	0.950	0.949	−0.18	−0.29	−0.11
RBF	tf-idf	Min, Lex & CUI	0.952	0.939††	0.945††	−1.32**	−0.66	0.67
RBF	tf-idf	Min	0.951	0.944††	0.952	−0.69	0.10	0.80
RBF	tf-idf	Min & CUI	0.949	0.941††	0.939††	−0.84	−1.06	−0.22
RBF	Count	Lex & CUI	0.949††	0.950†	0.953	0.04	0.42	0.38
Linear	tf-idf	Min, Lex & CUI	0.948†	0.930††	0.956	−1.99**	0.78	2.82**
RBF	Count	Lex	0.948††	0.950†	0.953	0.19	0.56	0.38
RBF	Count	Min	0.948††	0.947††	0.953	−0.04	0.59*	0.64
RBF	tf-idf	Lex & CUI	0.947††	0.946†	0.956	−0.08	0.96**	1.04
Linear	tf-idf	Lex & CUI	0.946††	0.942††	0.955	−0.40	0.93	1.33**
Linear	tf-idf	Min	0.944††	0.941††	0.952	−0.35	0.89	1.24**
Linear	Count	Min & Lex	0.943††	0.955	0.950†	1.23**	0.66	−0.57
Linear	Count	Lex & CUI	0.941††	0.944††	0.948††	0.30	0.78	0.47
RBF	tf-idf	CUI	0.938††	0.905††	0.933††	−3.51**	−0.53	3.08**
Linear	Count	Lex	0.938††	0.944††	0.946††	0.64	0.85	0.21
Linear	Count	Min, Lex & CUI	0.937††	0.955	0.950†	1.87**	1.29*	−0.57
Linear	tf-idf	Min & Lex	0.934††	0.948	0.951	1.52*	1.81**	0.29
Linear	Count	Min	0.934††	0.941††	0.950†	0.83	1.70**	0.86
Linear	Count	Min & CUI	0.931††	0.941††	0.950†	1.12	1.99**	0.86
RBF	tf-idf	Lex	0.929††	0.944†	0.937††	1.61**	0.80	−0.80
RBF	Count	CUI	0.928††	0.867††	0.929††	−6.66**	0.02	7.16**
Linear	tf-idf	CUI	0.924††	0.922††	0.929††	−0.26	0.55	0.82
RBF	Count	Min & CUI	0.923††	0.947††	0.953	2.56**	3.21**	0.64
Linear	tf-idf	Lex	0.921††	0.946†	0.941††	2.63**	2.12**	−0.50
Linear	tf-idf	Min & CUI	0.906††	0.941††	0.952	3.77**	5.01**	1.20*
Linear	Count	CUI	0.866††	0.861††	0.869††	−0.51	0.40	0.92*

^*percent change significant at alpha = 0.05

^**percent change significant at alpha = 0.1

^†statistically significantly different from highest performing model (emphasized above) at alpha = 0.05

^††statistically significantly different from highest performing model (emphasized above) at alpha =0.1

Note: For visual reference, models which are not statistically significantly different from the emphasized model at alpha = 0.1 are shaded in dark grey. Models which are not statistically significantly different from the emphasized model at alpha = 0.05 are shaded in light grey