Page 74 - Çevre Şehir İklim İngilizce - Sayı 4

Urban Transformation Forecasts with Artificial Intelligence
Based Algorithms

When Table 11 is examined, it is seen that the performance statistics of the
models established using the GWO algorithm give good results in verification
data sets, but from the point of view of training and test data sets, lean ANN
models exhibit better performances. When a general evaluation is made, it is
seen that ANN_GWO does not improve the performances of ANN models.
Of the models developed using different numbers of neurons, it has been
determined that the model with the highest accuracy is the ANN_5 model,
which was established using 5 neurons. When regression-based models and
ANN-based models are compared, it is seen that the model developed using
the TreeNet method shows higher performance compared to the ANN models.
In order to make this comparison clearer, time series of the number of licenses
variable have been prepared (Figure 5).

50000
Gerçek ruhsat Sayısı TreeNet ANN_5
45000
40000
35000
Ruhsat sayısı 25000
30000

20000
15000
10000
5000
0
1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76
Number of data
Figure 5. Time series of the number of licenses variable
When the prepared graph is examined, it is seen that the three peak values
included in the verification data set cannot be estimated with sufficient accuracy
in the model of both methods. In addition, it is observed that the estimated
number of licenses in the training and test data sets are close to the actual
number of licenses. The numbers of licenses obtained from the prediction
models and the actual numbers of licenses were compared by preparing scatter
graphs for training, verification and test data sets respectively (Figure 6).

Year 2 / Issue 4 / July 2023 63