Skip to main content

Table 2 Detection characteristics and parameters with different sliding window sizes and likelihood assumptions

From: Two-step spatiotemporal anomaly detection corrected for lag reporting time with application to real-time dengue surveillance in Thailand

Likelihood

Detection

Window size (weeks)

 

model

characteristic

5

10

15

20

25

30

 

Max accuracy

0.7143

0.7214

0.7175

0.7234

0.7209

0.7202

Poisson

Cut-off (percentile)

0.97

0.98

0.95

0.93

0.98

0.98

 

Time

0.5376

2.1435

6.5293

14.6618

28.5325

48.6852

 

Max accuracy

0.7081

0.7158

0.7145

0.7214

0.7227

0.7222

 

Cut-off (percentile)

0.93

0.95

0.94

0.95

0.98

0.98

Generalized

Time (min)

0.5487

2.2986

6.7418

15.8412

30.8942

53.2669

Poisson

Overdispersion delay

0.0862

0.0848

0.0937

0.0861

0.0923

0.0918

 

95% CrI

(0.041, 0.124)

(0.051, 0.135)

(0.055, 0.167)

(0.094, 0.191)

(0.092, 0.148)

(0.091, 0.129)

 

Overdispersion cluster

0.158

0.1636

0.1551

0.1534

0.1478

0.1466

 

95% CrI

(0.045, 0.374)

(0.046, 0.384)

(0.042, 0.371)

(0.042, 0.364)

(0.042, 0.353)

(0.041, 0.351)

Back to article page

BMC Medical Research Methodology

ISSN: 1471-2288

Contact us