Civil Engineering Reference
In-Depth Information
Tabl e 3
Bandwidths using
LCLS and LLLS regression
for region B during
2005-2010
Index
2SD
LCLS
LLLS
1
.
1783
×
10
2
z2
5.0414
14.9063
z3
4.3859
0.4821
0.9848
z4
3.9193
0.0421
1.5294
z6
5.8604
2.5685
1.6704
1
.
3407
×
10
2
z7
27.3238
47.8717
1
.
5100
×
10
4
3
.
8412
×
10
10
∗
z8
.
×
10
3
.
×
10
3
.
×
10
3
z9
6
6885
1
8810
2
6441
z10 14.6719 2.9917 4.6419
z11 2.5946 0.4757 1.0257
z12 8.4206 6.7895 3.0448
Notes: A bandwidth with a * next to it indicates that this variable
is smoothed out of the regression
Tabl e 4
Bandwidths using
LCLS and LLLS regression
for region C during
2005-2010
Index
2SD
LCLS
LLLS
10
7
∗
z2
54.2806
2
.
0978
×
z3
3.0987
0.4104
1.0937
10
3
∗
z4
3.3885
0.3080
7
.
7152
×
z6
5.1086
5.1488
2.3050
10
2
10
7
∗
z7
1
.
2713
×
5
.
2941
×
10
4
10
10
∗
z8
1
.
6882
×
1
.
0096
×
10
9
∗
z10 16.1393 5.8349 3.3952
z11 2.8887 0.4556 4.4160
z12 19.9572 3.6778 41.8698
Notes: A bandwidth with a * next to it indicates that this
variable is smoothed out of the regression
10
3
z9
6
.
0790
×
2
.
1734
×
4
Conclusion
Synthesizing all possible determinants of average medical expense from both
macro-economic and interior professional perspectives, the research establishes
a semi-parametric regression approach with LCLS and LLLS techniques to first
confirm effective determinants with relevance and further identify decisive and
control variables with linearity that contribute to average medical expense for
outpatients. The identification results are provided as follows.
In region A, average treatment number of outpatients for each physician per
day (z3) and illiteracy rate (z12) are decisive variables, with linear impact effects
on average medical expenses 0.00125 and 0.035, respectively. While number of
outpatients (z2), governmental health expenditure (z7), urbanization (z10), and
aging rate (z11) are control variables, which exert indirect impact on average
medical expense for outpatients.
Search WWH ::
Custom Search