Case study

 

1. Dataset

Daily hand-foot-mouth disease (HFMD) data from 19 hospitals in a district of China from August 4 to September 4 in 2009 and 2010, respectively, are provided as experimental historical data (see Fig. 1-1). The daily records of five hospitals in 2010 in the same study region are provided as sample data (see Fig. 1-2).

 

Fig. 1-1 Table of historical data

 

Fig. 1-2 Table of sample data

 

2. Exercise 1

Pop estimation

Step 1: To get started with the population estimation, first make sure that the inputted data format is correct. Please refer to the above figures to format the data (comma separated text file). Click 56D69}]YR02894}{MKAKVRH to add the sample, historical and result data files (see Fig. 2-1).

 

Fig. 2-1 Pop estimation

 

Step 2: Choose the total radio button to estimate the total/mean population, and then click ]H29@GM_@$GC0RTU8OX0I7B to calculate the model, as shown in Fig. 2-2.

 

 

Fig. 2-2 Load data

 

Step 3: The estimated total population is shown at the bottom text box in Fig. 2-3. 

Fig. 2-3 Computed total population and model parameters

 

 

3. Exercise 2

Samples selection with fixed number

Step 1: We continue to use the experimental historical data. Input the historical data file, give the required number n of sample stations (Fig. 3-1). Click the “Options” button to set the advanced options about the adoped simulated annealing algorithm, as shown in Fig. 3-2. 

 

Fig. 3-1 Samples selection

 

 

Fig. 3-2 Options

Step 2: The best combination in the simulation is shown in Fig. 3-3. Among all combination outcomes, the combination with the smallest estimated variance is selected as the best sampling choice, whose standard deviation s is the least of all combinations of 5 hospitals.

 

Fig. 3-3 Best combination of 5 stations in total population

 

 

Samples selection within a range

We can select samples with a batch mode that several sample schemes can be outputed at same time (Fig. 3-4). User can select the optimal scheme by considering both sample number and theoretical variance (V-N plot, Fig. 3-5).

Fig. 3-4 Samples selection in batch model

 

 

Fig. 3-5 V-N plot