Developing a portable natural language processing based phenotyping system

Table 10 The classification results for major classes without family history on the 15 types of selected CUIs

	P-Micro	P-Macro	R-Micro	R-Macro	F-Micro	F-Macro
Intuitive
LR	0.9001	0.6695	0.9001	0.5979	0.9001	0.6206
SVM	0.9074	0.6725	0.9074	0.6065	0.9074	0.6274
DT	0.9285	0.6783	0.9285	0.6355	0.9285	0.6467
RF	0.8690	0.6417	0.8690	0.5740	0.8690	0.5952
Textual
LR	0.9188	0.8037	0.9188	0.7303	0.9188	0.7608
SVM	0.9273	0.8060	0.9273	0.7388	0.9273	0.7669
DT	0.9538	0.8160	0.9538	0.7633	0.9538	0.7849
RF	0.8864	0.7823	0.8864	0.7081	0.8864	0.7386

For each task, the best results are bolded. The underlined results can achieve the top 10 results reported in [20]

ISSN: 1472-6947