This is my thesis project for my conversion MSc course in Informatics at Uni. of Piraeus 2021.
The code was deeloped in Python. The aim is to develop and test a MLP neural network model and a Support vector machine model to identify the sphere size distribution of materials based on 1D NMRI k-space data. Simulated data are used in this project.
This project's main idea is based on my previous PhD work. Relevant publication can be found here: https://doi.org/10.1016/j.jcis.2015.09.066
The previous work used a custom Bayesian model to clasify 1D NMRI signal. In contrast to that I know use MLP and SVM models instead. Moreover in this current work we approach the problem as a regression problem not as a classification one. Therefore the MLP and SVM models developed here are regression models.