hul 2004 International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part BS. Istanbul 2004 
  
3.2 Efficient Criteria and Factors 
By checking study area properties, available data layers and 
industrial conditions for industrial estates site selection, criteria 
and factors are classified in four main groups as below. 
Accessing Factors: Layers consist of Freeway, Highway, 
Asphalt Road 1, Asphalt Road 2, Asphalt Road 3, Gravel Road, 
Railway, City, Educational Centre, Medical Centre, Police 
Station, Train Station, Village and Airport. 
Infrastructure Parameters: Layers consist of Power 
Transformation Line, Water Supply, Oil and Gas Station, 
Telephone, Telegraph and Post Office. 
Environment Parameters: Layers consist of Urban Area, 
Forest and Green Space. 
Natural Factors: Layers consist of Slope Map, River and Wind 
Direction. 
3.3 Data Preparation 
tors and Before entering data into analytical models, they must be 
density prepared with respect to models execution routine and required 
ctors are inputs used by processing methods. There are many data 
case of processing methods in GIS environment, but we selected four of 
political them such as Layers Combination, Data Structure Conversion, 
s better Distance Map Supplying and Data Classifying. Major 
parameter that affects on processing methods selection is data 
such as input structure (Bonham Carter and G.F., 1991). Raster 
ind land structure was selected as input structure in our application 
ts' ideas because of its properties such as simplicity and calculation 
factors, ability for integrating layers, and mentioned models (Section 2) 
of Iran. execute routine. In here, there is a major point called the 
arca and optimum data volume capacity that with respect to the scale of 
rovince) maps and required accuracy can be determined. Required 
here are accuracy in our application was determined to be 0.3 millimeter 
on the map scale using experts ideas and required content map. 
Therefore, maximum pixel size will be 7.5 * 7.5 meter on the 
ground. 
Accordingly, data processing methods were applied on raw data 
and resulted factor maps. Each pixel were applied on the factor 
Ji on maps has a gray value that indicates amount of proportionality 
for industrial estate construction with comparing other pixels 
value or spatial units. For producing factor maps we must do 
regular steps using processing methods that depend on required 
criteria and factors. For example there is a vector structure 
Wind Direction layer of Natural class that must be converted to 
raster form using Data conversion method and classified based 
on spatial data priority using Data Classifying method. This 
priority is determined based on expert ideas and spatial units 
value. In the wind direction layer, spatial units that their 
direction are toward city, take less gray valuc than others 
because of controlling air pollution in the city. 
e of 
er of 3.4 Data Weighting 
versity 
The weight of each factor map indicates amount of its cost and 
value as comparing with the other factor maps. Correct weights 
can help finding convenient location for industrial estate area. 
There are two ways for weighting factor maps that are called 
281 
Knowledge Driven and Data Driven Weighting. 
Knowledge Driven Weighting: In this method, data are 
weighted in a definite range using expert experience about 
application. First, different ideas are collected and their 
dimensions are uniformed. Then, weights are normalized in the 
defined range using an appropriate scale. 
Data Driven Weighting : Weight of each factor map will be 
determined by calculating amount of dependency value 
between factor maps and specified sites layer. One of the 
appropriate methods for determining dependency value is 
Weight of Evidence. This is a probabilistic method that can be 
used for integration of factor maps. This model is executed 
based on probability theory and uses preferent and recent 
probability rules similar to Baysian model (Bonham Carter and 
G.F., 1991). There are two parameters in the weight of evidence 
method that are named sufficiency and necessity ratio. They 
specify amount of cach factor map sufficiency and necessity for 
industrial estates location using follow Equation. 
p DON ror ad using ENSee a 
P(B; | D) P(B; | D) 
where B;= The effective zone in cach factor 
"B= The supplemental zone of B; 
D- The existent sites location 
LN- Necessity ratio 
LS- Sufficiency ratio 
Then, by Naperian logarithm computing LN and LS, positive 
weight of evidence and negative weight of evidence for each 
factor map can be determined (Bonham Carter and G.F., 1991). 
Finally, by computing constant parameter that called factor map 
contrast, dependency value is determined (Equation 9). 
C=|W,-W._|=[Ln(LN)-Ln(LS)| (9) 
Where — W,- positive weight of evidence 
W = negative weight of evidence 
Ln= Naperian logarithm 
C= factor map contrast 
Table 1. Weighting of factor maps 
  
  
  
  
  
  
  
    
   
  
  
  
  
  
  
   
   
   
  
  
  
  
  
   
    
    
   
     
    
    
   
    
    
   
    
  
      
   
   
   
    
    
   
  
  
  
  
  
  
  
    
  
  
   
  
  
    
  
   
  
  
  
  
  
  
    
  
  
    
  
  
   
   
  
  
Class Factors Weight Class Factors Weight 
Main Roads 0.10 Forest (0.35 
Environment 
Asphalt Parameters E 
Green Area 2 
Roads 0.09 0.20 reen Área 0. 
Gravel 0.07 City area 0.45 
Road 
ailwav 2 S » M: A 
Railway 0.12 Natural Slope Map 0.45 
City 0.09 Factors River 0.20 
Education 0.07 024 Wind 0.35 
Centre Direction 
Access Medical am Tr 95 
Factors Centre 0.07 Power Line 0.25 
Vater > 
0.27 Airport 0.13 Walel (0.25 
Resource 
Post 
Police Infrastructure Telegraph 
ur 0.06 Parameters and 0.25 
Station 5 Telent 
0.29 elephone 
Office 
Train 
Station 0.10 Oil and Gas 0.25 
Station = 
Village 0.10