Optimal Decision Trees for Algorithm Selection
Research on dynamic-programming methods for learning optimal decision trees.
This project explored how optimal decision trees can be used for algorithm selection problems, where the goal is to choose the best algorithm for an instance based on its features.
The work involved dynamic programming, empirical evaluation, and implementations in C++, Java, Python, and R. It grew out of my B.Sc. thesis and follow-up research assistant work at Delft University of Technology.