Optimising Class-Imbalanced Data Identification in Financial Anti-Fraud through Reinforcement Learning
This project focusses on addressing the class-imbalanced data issue in financial anti-fraud areas. Students are tasked to apply data preprocessing models and classifiers to identify anomalous samples across multiple datasets, working towards effective solutions for optimisation challenges and enhancing model performance. In addition, students are demanded to explore the utilisation of reinforcement learning to guide symbolic regression in discovering the optimal components of interpretable models.