Optimal Decision Tree

Optimal decision tree (ODT) research focuses on developing algorithms that globally optimize tree structure for improved classification or regression accuracy, unlike traditional greedy approaches. Current research explores various objective functions for optimization, efficient algorithms like mixed-integer linear optimization and branch-and-bound methods, and techniques to handle noisy data and distribution shifts. This work aims to improve the accuracy and robustness of decision trees while maintaining or enhancing interpretability, impacting both machine learning methodology and practical applications requiring transparent and accurate predictive models.