62
H. Altıparmak et al.
Fig. 9.2 Neural network performance
With this application, we have seen that we can predict future climate change
situations with neural networks that produce more successful results. As can be seen
from the database we used, all the factors that make up the atmosphere can have an
impact on climate change. As a result, we used a dataset consisting of MEI, CO2,
CH4, N2O, CFC-11, CFC-12, TSI and Aerosols. It is observed that our training
success is 90.172%, validation success is 84.859%, test success is 81.697% and
overall success is 87.945% (Fig. 9.2; Table 9.1).