论文部分内容阅读
目的对广东省时空模型试点县区在预警阈值调整前后的时空模型和时间模型的预警效能进行评价,了解时空模型的预警效能是否更优。方法对2008年4月21日-2012年12月31日广东省时空模型试点县区的时空模型和时间模型的灵敏度、特异度、及时性和热点定位进行评价。结果预警阈值调整前,时空模型与时间模型的灵敏度分别为67.74%和35.48%(χ2=12.917,P<0.01),特异度分别为20.25%和8.23%(χ2=28.063,P<0.01),及时性M分别为-0.49 d和1.65 d;预警阈值调整后,时空模型与时间模型的灵敏度分别为78.43%和43.14%(χ2=13.326,P<0.01),特异度分别为31.82%和13.35%(χ2=34.325,P<0.01),及时性M分别为-0.27 d和0.67 d;预警阈值调整前后,时间模型均无法从空间上识别热点区域,而时空模型则可以发现空间聚集性异常。结论无论是预警阈值调整前还是后,时空模型的灵敏度、特异度、及时性和热点定位均优于时间模型;预警阈值调整后的时空模型的灵敏度、特异度、及时性和热点定位均优于调整前。
Objective To evaluate the early-warning efficiency of time-space model and time model before and after the pre-warning threshold adjustment in the pilot counties of Guangdong Province spatio-temporal model to find out whether the pre-warning efficiency of the spatio-temporal model is better. Methods The sensitivity, specificity, timeliness and hot location of spatio-temporal model and time model of Guangdong province spatio-temporal model pilot counties from April 21, 2008 to December 31, 2012 were evaluated. The sensitivity of spatio-temporal model and time model were 67.74% and 35.48% (χ2 = 12.917, P <0.01) before the adjustment of pre-warning threshold. The specificity was 20.25% and 8.23% respectively (χ2 = 28.063, P <0.01) The sensitivity of spatio-temporal model and time model were 78.43% and 43.14% respectively (χ2 = 13.326, P <0.01) with the specificity of 31.82% and 13.35%, respectively χ2 = 34.325, P <0.01). The timeliness M was -0.27 d and 0.67 d, respectively. Before and after the warning threshold was adjusted, the time models could not identify the hot spots spatially, but the spatio-temporal models could find the spatial aggregation abnormalities. Conclusion The sensitivity, specificity, timeliness and hot spot location of spatio-temporal model are better than that of the time model before and after the pre-alarm threshold adjustment. The sensitivity, specificity, timeliness and hot spot location of the spatio-temporal model after pre-alarm threshold adjustment are better than before fixing.