The Practical Application of Decision Tree-Based AI Models

Abstract

From Chat-GPT to the algorithm behind your Netflix recommendations, a substantial portion of what we know as Artificial Intelligence begins life as a decision tree. In this interactive course, we will delve into the practical application of decision tree-based AI models. We are not going to explore deeply into the mathematics behind machine learning algorithms; however, we will recommend suitable resources for mathematically-minded participants. We'll commence with a comprehensive exploration of decision trees themselves (npart), then progress to more advanced techniques including Random Forests (Ranger) and sophisticated gradient boosted models like XGBoost. These machine learning models not only function as powerful statistical tools but also prove indispensable in a wide array of applications. The entire session will be conducted within the R environment. During the session, annotated scripts and data will be provided, making prior R experience optional. However, experienced R users are encouraged to bring along any additional datasets they'd like to explore. To ensure seamless participation, we kindly request that participants install the caret package in advance, as it may encounter compatibility issues with specific versions of Rtools. You won't need access to an extremely powerful computer to participate; for reference, I will be using a 2017 Macbook Air A1466. We hope to see you there.

Schedule

10:00-11:00 Introduction: Decision Trees, Random Forests & Extreme Gradient Boosting
11:00-11:15 Coffee break
11:15-12:15 Applying Machine Learning to Linguistic Problems
12:15-13:30 Lunch break
13:30-15:00 Applied Learning Session