abstract = "Genetic Programming (GP) is an evolutionary
computational method that can generate symbolic and
mathematical models. In addition to being a type of
evolutionary computing, the GP is also frequently used
in solving symbolic regression problems in machine
learning applications. Since its first appearance, it
has become one of the popular evolutionary calculation
methods as a result of being successfully applied for
solution of modeling problems appeared in many
different disciplines. Within the scope of this thesis,
research studies have been carried out for development
of data driven prediction models and their engineering
applications by using the classical GP and its a
variant, Gene Expression Programming (GEP). In order to
increase the effectiveness of these GP methods in
practice, data normalization, ensemble learning, hybrid
model development and hyperparameter optimization
techniques are studied. In addition, the chromosome
structure of the GEP method has been modified and an
optimal solution to the constant value determination
problem has been proposed. Then, the modified GEP
method was combined with popular metaheuristic
optimization methods, and thus a metaheuristic
optimization based GEP (MetaSezGEP) approach was
developed. Contributions of these improvements to some
engineering applications have been investigated.",