De Waele, Lena2025-08-292025-09-302025-01-16http://hdl.handle.net/10362/187176"Lanthanide-doped nanoparticles (NPs) have shown great potential in biomedical applications, particu- larly in imaging and targeted therapies, due to their unique photoluminescence (PL) properties (2,3). Recent advancements in machine learning (ML) offer new opportunities for optimising NP synthesis. ML enables algorithms to learn from data, potentially accelerating discoveries, but significant challenges remain (4–6). Additionally, genetic algorithms (GAs), inspired by natural selection, can efficiently search through complex parameter spaces, making them suitable for optimising synthesis conditions (7–9). This study explores the potential of artificial intelligence (AI), specifically combining ML and GAs, to model and optimise NaGdF4:Eu NP synthesis conditions, aiming to maximise crystallite sizes while maintaining a small particle size. Increasing crystallite size has been shown to enhance PL intensity in various doped NPs. This approach can accelerate biomedical advancements efficiently in terms of both time and resources. Our approach included feature selection, standardisation, and one-hot encoding to build robust models.(...)"engNanoparticles,Genetic AlgorithmsCrystallisationMachine LearningOptimisationmaster thesis