Automatic artificial neural network model development tool for dataset

Artificial neural networks are one of the most important and groundbreaking machine learning algorithms in the field of artificial intelligence. Artificial neural network models work behind many important and popular artificial intelligence applications today. However, the development of an artificial neural network model that can perform a specific task on a specific data set is based on the knowledge and experience of a human expert. There are various studies in the literature that use different techniques to automate this human-made process. This thesis study aims to develop a software that automatically develops artificial neural network models. For this purpose, a desktop application software called Automatic Artificial Neural Network Model Development Tool for Dataset has been developed, which automatically develops the best artificial neural network model that can perform tasks such as regression and classification for a specific data set. The architectural structure of the artificial neural network models automatically developed by this software is in the form of multilayer perceptron, which is the classical artificial neural network architecture. The automatic model development process of this software is carried out by an optimization algorithm called genetic algorithms. Genetic algorithms automatically enable the determination of the most appropriate layer parameters and the most appropriate hyperparameters of artificial neural network models through the evolutionary genetic processes they provide. Users using this software can control all these processes through the desktop application interface and see the search results. They can even export the model that is the best result of the search in a re-executable software format.