Mohamed Lafif TEJ, Public Dissertation of PhD Thesis

Thesis Title: ”Contributions to the Determination of Neural Network Architectures PhD thesis”

PhD Board:

• Chair: Professor Dr. Eng.  Marius MARCU  (Politehnica University Timisoara)
• PhD Supervisor: Professor Dr. Eng. Ștefan HOLBAN (Politehnica University Timisoara)
• Scientific Referees:
  • Professor Dr. Eng. Cornelia Aurora GYŐRÖDI (Oradea University)
  • Professor Dr. Eng. Alexandru CICORTAȘ  (Vest University of Timisoara)
  • Professor Dr. Eng. Ionel JIAN (Politehnica University Timisoara)

The thesis aims to create a general method to determine the optimal neural network architecture using data mining techniques. Data mining are used to explore training data of a neural network to identify patterns and establish relationships that can lead to reasonable predictions of the optimal architecture of a multi-layer neural network.This thesis presents a new general method to determine the optimal neural network architecture using Linear Regression. Starting from clustering the dataset used to train a neural network it is possible to define Multiple Linear Regression models to determine the architecture of a neural network. This method is more flexible with different datasets types unlike other methods. The proposed method adapts to the complexity of training datasets to provide the best results regardless of the size and type of dataset. The proposed method reduces the time allocated for network design. This method can make the design simple and easy and possible for a Non-specialist designer. The focus was to develop a fast solution (in terms of learning iterations) maintaining also an acceptable efficiency