Geography Reference
In-Depth Information
3
3
ð
s
ij
¼ t arg et
Þ
X
ij
¼
3
3
1
con s
ij
is unit constraint condition represent cellular which cannot transform to
urban land, such as water and high mountain. For example, when a cell represents
basic farmland preservation area, con s
ij
0, because it is constraint to devel-
opment in this area.
P
g
is regional land use change regulations mined from data of regional land use
and related influencing factors. In our research, C5.0 Decision Tree Algorithm is
applied to calculate regional land use change rules. Decision Tree Algorithm is a
typical data mining classification algorithm. Its main role is to reveal the structured
information of the data. The created tree structure is visual, easy to understand, and
deal with nonlinear data. The hidden knowledge rules in data can also be extracted.
Therefore, Decision Tree Algorithm can be used to dig out cellular transformation
rule (Ke et al.
2009
). In C5.0 Tree Decision Algorithm, clusters are determined by
the fellow formula.
I
ð
r
1
;
r
2
;
...
;
r
m
Þ
¼
X
m
p
i
log
2
ð
p
i
Þ
i¼1
where r
i
is the subset of dataset S which belongs to cluster C
i
, P
i
is the probability
of every sample belongs to C
i
, and I is the information gain.
Asynchronous evolving speed is determined by two parts, level of social-eco-
nomic development and urbanization scenario. Compared with the conditions of
social and economic development level, urbanization patterns play much more
important role in urban land expansion speed. Therefore, in order to clarify the
influential difference of urban land expansion on regional climate change in dif-
ferent urbanization pattern of Wuhan Metropolitan, asynchronous evolving speed
for Partitioned and Asynchronous Cellular Automata Model is decided by
urbanization patterns. Under baseline scenario, the evolving speed for each cellular
in Wuhan Metropolitan follows in the history of law, mainly determined by the
regional differences of socioeconomic development. Under centralized urbaniza-
tion scenario, the greater the cities are, the higher the urban land expansion prefers
and the faster urban land expands. Under decentralized urbanization scenario, the
smaller the size of cities is, the higher priority and the slower speed of urban land
expansion is. Asynchronous evolving speed in Partitioned and Asynchronous
Cellular Automata Model can be figured out by the following formula.
priority
ij
priority
max
priority
min
v
ij
pri
¼
v
max
v
min
ð
Þ þ
v
min
Where v
ij
is transformation speed of cell (i, j), priority
ij
is urban land devel-
opment priority of cell (i, j), priority
max
is the maximum of all-region's urban land
