Geoscience Reference
In-Depth Information
Table 4
Results obtained from IDW method
Inverse distance power
RMSE from first series (m)
RMSE from second series
(m)
CHC (1)
CHC (2)
CHC (1)
CHC (2)
1
2.940
3.489
0.902
0.956
2
2.003
2.417
0.779
0.854
3
1.565
1.859
0.693
0.792
The second part is the independent check points, is used to evaluate the final chro-
mosome, known as Independent Check (ICPs). In this paper series of check points
(1) and (2) are as GACPs and ICPs respectively.
7.2 Using Genetic Algorithms in Optimisation of Inverse
Distance Weighting Method
GA, control and check points can be used to optimise the magnitude of weight and
consequently proper weight is achieved. In this article, from the control points and
the series of checkpoints (1) as the GACPs can be used to evaluate the strength
of optimisation. Finally, the strength obtained from GA is substituted in the IDW
equation and as a result the accuracy of the algorithm from the series of check-
points (2) as the independent checkpoints obtained through the optimisation pro-
cess with GA (ICPs) are examined and evaluated.
7.3 Assessment of Results
IDW interpolation method is used for changing the power of inverse distance and
both data sets have been tested and presented in Table
4
.
As shown in Table
4
, the RMSE rate with different powers of inverse distance
on two sets of series, the checkpoint values is obtained in metres. In both data sets
minimum RMSE values are obtain, which is greater to the power of 3.
In Kriging method, the interpolation of numerous Vary-grams, such as; spheri-
cal, linear, exponential, Gaussian, in the position no drift, linear drift and quadratic
drift are used and the result is compared in Table
5
.
As shown in Table
5
, the accuracy of the Gaussian and spherical Vary-gram
equivalent and the highest accuracy are achieved using linear Vary-gram.
In all cases, linear and quadratic drifts are more accurate compared to no drift,
but there is not much difference between quadratic and linear drift. Results of
other methods have been shown in Table
6
.
Using GA, control points and checkpoints (1) can be optimised for each set of
examined test points that are obtained. The results can be checked against check-
