Geoscience Reference
In-Depth Information
Table 5
Results from Kriging method
Type of vary-gram
Type of drift
RMSE from first series (m)
RMSE from second
series (m)
CHC (1)
CHC (2)
CHC (1)
CHC (2)
Spherical
No drift
4.020
4.646
1.002
1.043
Linear drift
2.194
2.351
0.943
0.999
Quadratic drift
2.193
2.350
0.932
0.989
Exponential
No drift
1.869
2.251
0.732
0.821
Linear drift
1.783
1.953
0.695
0.796
Quadratic drift
1.782
1.952
0.690
0.793
Linear
No drift
1.242
1.361
0.581
0.728
Linear drift
1.309
1.391
0.574
0.725
Quadratic drift
1.309
1.390
0.573
0.724
Gaussian
No drift
4.020
4.646
1.002
1.043
Linear drift
2.194
2.351
0.944
0.999
Quadratic drift
2.193
2.350
0.933
0.990
Table 6
Results obtained from other conventional interpolation methods
Interpolation method
RMSE from first series (m)
RMSE from second series (m)
CHC (1)
CHC (2)
CHC (1)
CHC (2)
Natural neighbour
1.782
1.952
0.933
0.990
Nearest neighbour
1.205
1.348
0.555
0.716
Triangulation
5.413
4.659
0.578
0.727
Quartic polynomial
1.429
1.501
0.484
0.684
Table 7
Results from IDW optimisation method with GA
Indices optimisation
First series checkpoints
Second series checkpoints
2.67
2.964
RMSE (m)
GACP
ICP
GACP
GACP
0.662
0.978
0.466
0.684
Table 8
Results of quartic polynomial optimisation with GA
Power of polynomials
No. of variables
RMSE from first
series (m)
RMSE from
second series (m)
GACP
ICP
GACP
GACP
4
15
0.662
0.978
0.466
0.684
polynomial containing 15 variables has been optimised with respect to GA and
the coefficients have been optimised and the proper terms extracted. In this case,
the optimisation methods like IDW has a set of control points and a series of
checkpoints (1) in the algorithm process used to extract the relevant terms and the
results are collected and evaluated against the checkpoints (2). Tables
7
and
8
compare the results that are achieved by using GA with the IDW method and
Quartic Polynomial optimisation respectively.
