Page 56 - Çevre Şehir İklim İngilizce - Sayı 4

URBAN TRANSFORMATION FORECASTS WITH
ARTIFICIAL INTELLIGENCE BASED ALGORITHMS

Hacı Abdullah UÇAN - Tayfun DEDE - Sinan NACAR

ABSTRACT
Urban transformation is a design practice at the focus point of modern urban
life. This phenomenon provides the purpose of human life improvement by
organizing it on an urban-based. Urban areas that were established and
developed depending on time, geographical features, and historical processes
usually have built against the engineering approaches, especially for economic
excuses and follow an uncontrolled and inadequate course, and progress in a
physical, social, and economic environment which is inappropriate for human
life. To improve such cities and to design future cities that include sustainable
structures, urban transformation interventions are becoming inevitable. In
addition, the risk factor caused by millions of inappropriate buildings against
disaster risk is an essential topic that should take place in the concept of urban
transformation planning and engineering studies in Türkiye. Considering this
essential reason that urban transformation practices can affect the economic
indicators of the future of Türkiye, it is emphasized that it is quite important to
make very realistic predictions. Using the periodic economic data of the urban
transformation applications in our country, estimation models were created for
the urban transformation process using regression and artificial intelligence-
based algorithms. Thus, by processing the predicted data as input, it will be
possible to make predictions of the urban transformation process in Türkiye. In
general, the ANN method gave better estimates than the traditional regression
methods. When the prediction models are examined in detail for the training,
validation and test data sets, some models obtained by adding gray wolf
optimization to the ANN yielded better results than the models established
with other methods. The best estimation model was obtained by using the
TreeNet method, with NS values of 0.62 for the number of flats, 0.70 for the
number of licenses and 0.67 for the surface area, according to the average
evaluation for the training, test and validation sets.

Keywords: Urban transformation, Regression Analysis, Artificial
Intelligence-Based Algorithms, Optimization, Prediction Models

Year 2 / Issue 4 / July 2023 45