TY - THES ID - 136725156 TI - Classification of Solar Energetic Activity using Data Analysis and Clustering of Active Regions AU - Baeke, Hanne AU - Lapenta, Giovanni AU - Amaya, Jorge AU - KU Leuven. Faculteit Wetenschappen. Opleiding Master of Astronomy and Astrophysics (Leuven) PY - 2021 PB - Leuven KU Leuven. Faculteit Wetenschappen DB - UniCat UR - https://www.unicat.be/uniCat?func=search&query=sysid:136725156 AB - Although the Sun looks to us peaceful and quiet when seen from afar, it is in reality a hot ball of gas that is in continuous change. The Sun is constantly boiling, bringing its internal energy to the surface in the form of radiation, mass and magnetism. Often, this energy ruptures the top layers of the Sun, in this way creating sunspots, which can be observed from Earth with telescopes. This energy is constantly released into space and can reach our planet. It is this energy that causes the beautiful auroras at the poles. It does not affect the people on Earth, because we are partially protected by the natural magnetic field of the Earth, but on occasions, when the solar activity is intense, our technological structure is at risk. In the most extreme case, this could knock out our electrical power grids, which would have catastrophic results in our modern society, causing considerable economic damage. Therefore, the study of the energetic activity of our Sun is very important. In this master thesis the properties of the solar magnetic field are studied at highly energetic regions on the Sun, since those two are strongly coupled. Also the relation with the amount of energy they release into space is studied. Machine learning is used for this research. Machine learning is a part of Artificial Intelligence and is the study of computer algorithms that improve through experience by the use of data, without being explicitly programmed to perform certain tasks. In this way, machine learning methods are able to find patterns in the data that might escape the human eye. For this research, these methods are used to let the computers detect by themselves the properties of the regions where large energy releases occur. As such, the data is categorized into regions which release more or less energy. Before analysing the results obtained with machine learning, the data is manually inspected. This learns us that the magnetic field of highly energetic regions on the Sun is very similar when energy is released. There is not much difference detected based on the amount of energy released. Therefore, it will be hard to distinguish them. However, there are also energetic regions that are not releasing energy at the moment of the measurement. These regions do have different properties and are therefore expected to be distinguishable. To conclude, the results obtained in this research learn us that it is possible to distinguish between the energy releasing regions and the regions that do not release energy. Moreover, the most energetic releases can sometimes also be distinguished from the less energetic releases. More accurate methods could be developed by combining this with other data, like for example the evolution in time or the shape of the highly energetic regions. ER -