The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part Bl. Beijing 2008 
987 
same meaning as the scan circle in spatial scan statistics. The 
height and radius values of scan cylinder also have specified 
maximum values. 
Suppose c zd is the observed number of cases in region z during 
day d. The total number of observed cases (C) is: 
C = I2X (2) 
For each region and each day, the expected number of 
cases n is calculated as follows: 
results of numerous disease cases. The diagnosis result is more 
precise with larger amount of sample cases. 
Two basic discriminatory analysis methods used in intelligent 
disease diagnosis are Fisher and Bayesian analysis. Bayesian 
analysis method which is based on probability and statistics is 
of higher preciseness and effectiveness. According to the prior 
probability of some certain event, its corresponding posterior 
probability could be calculated by Bayesian analysis. 
Suppose eventsA,,A 2 ,A 3 ,...A n comprise a complete event group. 
For each event j,, its occurrence probability is defined as 
P(A j ) (P(A ) > 0,/ = 1,2,...,« )• Under the condition of event# 
occurring, for each events., its occurrence probability 
P(A i | B) could be calculated as the following expression: 
(3) 
Under the null hypothesis, the expected case number of 
particular scan cylinder A could be expressed as follows: 
P{A,\B) = 
P(A i )*P(B\A i ) 
P(B) 
P(A i )*P(B\A i ) 
¿P(4)*P(#I4) 
(6) 
Pa = YjPtd 
(z,d)eA 
(4) 
Suppose C A is the observed number of cases in a particular 
cylinder A. When y c and y ^ are small compared 
to C , C 4 is approximately Poisson distributed with mean 
(Evans, 2000). Poisson generalized likelihood ratio (GLR) is 
used as a measure of the evidence that cylinder A contains a 
disease outbreak (Kulldorff, et al, 2005). GLR value could be 
calculated as follows: 
GLR = 
(c ) 
C, 
fc-o 
UJ 
S' ~ Pa ; 
(5) 
3.2 Bayesian analysis 
One important module of epidemical response system is 
intelligent disease diagnosis. According to the epidemical 
fieldwork investigation, clinical examination of patients and 
examination results of laboratory, intelligent disease diagnosis 
module uses specified algorithm and program to complete quick 
judgement to the bursting case and diagnosis to unknown 
diseases. Epidemical response personnel can choose specified 
disease from the diagnosis result on the basis of their 
experience or expert suggestions, and the system could provide 
relevant response suggestions according to the diagnosed 
disease. 
Intelligent disease diagnosis uses possibility as the judgement 
measure, and proposes the diagnosis result according to the 
information of symptoms, disease characteristics, laboratory 
examination, etc. The relationship between disease classes and 
judgement information should be determined from analysis 
For the application in intelligent disease diagnosis, suppose the 
patient has « kinds of clinical symptoms corresponding to m 
kinds of diseases. For disease j (j = \,2,...m) corresponding to 
symptom z (/ = 1,2,...,«), the probability is defined as/>(#.,^4 ). 
Meanwhile, the incidence rate of certain disease j is defined 
as P(Aj) ■ Therefore, the occurrence probability of disease^ 
corresponding to all the « kinds of clinical symptoms could be 
calculated as follows: 
P(A\B) WWAj) P(A j )*P(B\A J ) 
P{B) f j P(A J )*P(B\A J ) (7) 
7=1 
P(A j )*f\P(B. n A J ) 
_ /=] 
X(P(zt,.)*n^P(B„ A.)) 
j=i m 
4. SYSTEM STRUCTURE DESIGN 
This collaborative system is designed in C/S architecture mode, 
and consists of collaborative server, PC client, field/PDA 
(Pocket Digital Assistant) client, and communication network. 
The people working in the system include server managers, PC 
client operators, and filed working people. The communication 
network comprises internet and mobile network. 
The structure design of collaborative system is shown in Figure 
1. The surveillance subsystem includes server and PC client, 
and the response subsystem includes server, PC client and PDA 
client. These two subsystems have the same server, and 
complete their collaborative work through PC client. 
The work flow is as follows: The server runs a full-time 
epidemical surveillance and detection program. Once an