Applying machine learning and optimization to high-throughput experimentation

• Main Author: Low, Andre Kai Yuan
• Other Authors: Kedar Hippalgaonkar
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2021
• Subjects: Engineering::Materials; Engineering::Computer science and engineering::Computing methodologies::Simulation and modeling
• Online Access: https://hdl.handle.net/10356/147754
• Institution: Nanyang Technological University

This study investigates the effects of different hyperparameters and settings in performance of Bayesian and Particle Swarm Optimization, and in doing so develop best practices for their practical application. Concrete Compressive Strength Data Set was modelled with Xtreme Gradient Boosting Regressor, with results being compared against a benchmark of Random Search. Additionally, various notebooks and code was developed to provide a resource for implementing machine learning in high-throughput experiments, and designing appropriate virtual experiments through the Monte Carlo method with Latin sampling. Different plotting functions were also developed to provide better visualization of the entire process.