Page 64 - Çevre Şehir İklim İngilizce - Sayı 4

Urban Transformation Forecasts with Artificial Intelligence
Based Algorithms
Eğitim veri seti (a)
Training data set (a)
600000 120000
Daire sayısı Ruhsat Sayısı Yüz Ölçümü
Surface Area
Number of
500000 100000
Number of flats Daire sayısı 400000 80000 Ruhsat sayısı/Yüz ölçümü Number of licenses /
300000
60000
200000
100000 40000 Surface area
20000
0 0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45
Doğrulama veri seti (b)
Verification data set (b)
350000 80000
Daire sayısı Number of Yüz Ölçümü
Ruhsat Sayısı
Surface Area
Number of flats
300000 60000
Number of flats Daire sayısı 200000 40000 Ruhsat sayısı/Yüz ölçümü Number of licenses /
250000
150000
100000
50000 20000 Surface area
0 0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Test veri seti (c)
Test data set (c)
350000 70000
Daire sayısı Ruhsat Sayısı Yüz Ölçümü
Number of
Number of
Surface Area
300000 60000
Number of flats Daire sayısı 200000 40000 Ruhsat sayısı/Yüz ölçümü Number of licenses /
50000
250000
30000
150000
100000
10000
50000 20000 Surface area
0 0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Figure 1. The time series of the number of flats, number of licenses and surface area
variables belonging to the training (a), verification (b) and test (c) data sets for the mixed
selection scenario

ANN-GWO Model

GWO or other algorithms optimize the weights in the training of ANN
weights, that is, they find the most appropriate values. The main goal in the
optimization of ANN is to define the optimal weight values and network
architecture. The diagram representing the collaboration of ANN and GWO
optimization technique is given in Figure 2. This model, which is established
for prediction models, is written in the MATLAB programming language.

Year 2 / Issue 4 / July 2023 53