Skip to main content

View full-text article in PMC

. 2018 Sep 25;4(3):e65. doi: 10.2196/publichealth.8627

Table 2.

Values of the correlation coefficient (r) of methods with natural language processing.

Target and method		SEASON 2012	SEASON 2013	SEASON 2014	SEASON-TOTAL
All areas
	TRAP+NLP^a	.76^b	.70^b	.69^b	.70^b
	LINEAR+NLP	.70	.55	.53	.50
	BASELINE+PROF^c+NLP	.74^d	.68	.67	.69
	BASELINE+NLP	.33	.37	.48	.36
High-population areas (Top 10)
	TRAP+NLP	.80^b	.77^b	.72^b	.75^b
	LINEAR+NLP	.78^d	.65	.64	.64
	BASELINE+PROF+NLP	.80^b	.77^b	.71^d	.75^b
	BASELINE+NLP	.55	.60	.63	.53
Low-population areas (Top 10)
	TRAP+NLP	.75^b	.66^b	.71^b	.69^b
	LINEAR+NLP	.62	.46	.48	.43
	BASELINE+PROF+NLP	.70	.61	.65	.64
	BASELINE+NLP	.21	.26	.35	.25

^aNLP: natural language processing.

^bHighest correlation coefficient in each target area and each SEASON.

^cPROF: profile information.

^dHigh correlation (r>.7).